Two-fan coolers have popped up here and there, but the NH-D14 is exceptional in being the flagship model of a manufacturer that specializes only in high performance CPU heatsinks. We sink our teeth deeply into this big feast from Noctua.
February 2, 2010 by Mike Chin
| Product | Noctua NH-D14 CPU Cooler |
| Manufacturer | Noctua |
| Street Price | ~$90 USD |
The NH-D14 is only Noctua’s fourth heatsink design since the company’s
emergence in 2005. The others are a 12cm fan tower, a 9cm fan tower, and a 12cm
fan top-down model. There are several variants of each and there have been subtle
refinements over the years, but those three basic designs have made up the line.
The low model count makes Noctua one of the most conservative heatsink specialists
in the computer business, and in this case, the word conservative is meant positively.
The company has also been diligent in updating the mounting systems of their
variant models to be compatible for changes by Intel and AMD, with update kits
made available for purchase, not only incorporated into new production. This
kind of customer support is unusual and we wholeheartely applaud Noctua’s approach.
The company’s products have always shown a high degree of systems design thinking
and very good attention to details. No surprise that they’ve received positive
reviews from many web tech reviewers, including SPCR. In our recent retest of
top heatsinks on our new i7 test platform, the Noctua
NH-U12P tower heatsink’s position among the top performers was re-established.
The NH-D14 was introduced late last year, just a couple of months ago. It is
considerably larger than their previous tower models, and looks designed to
handle the hottest CPUs around today — the Intel Core i7-1366 (130W TDP)
and the AMD Phenom II 965 (125W~140W TDP).
|
|
The large twin-tower design features two fans, a 14cm in between the two banks
of fins, and a 12cm fan blowing in the same direction on the outside of one
of the banks. The overall presentation of the product itself as well as the
neat packages of Intel, AMD and common hardware cannot fail to impress.
|
|
| Noctua NH-U12P: Key Features (from the product web page) | |
| Feature & Brief | Our Comment |
| 6 heatpipe dual radiator design Providing more surface area, better heat-distribution and superior airflow efficiency than conventional tower style heatsinks, the NH-D14’s six heatpipe dual radiator design was developed to provide ultimate quiet cooling performance in dual fan mode. | Because they are U-shaped, these 6 heatpipes provide 12 thermal paths. We’ve seen twin-tower designs before but haven’t been that impressed. |
| Dual NF-P14/NF-P12 fan setup Premium quality dual fan setup consisting of Noctua’s award-winning NF-P12 (120mm) and NF-P14 (140mm) fans, both of which feature Vortex-Control Notches, SCD technology and SSO-Bearings in order to achieve a perfect balance of performance and quietness. | Two fans blowing in the same direction should provide less drop in pressure as RPM is reduced, which means better cooling at low (read: quiet) speed. But turbulence interactions between the fans can increase noise. We cannot confirm the utility of the VCN, SCD technology or SSO bearings. |
| Asymmetrical design for high compatibility An asymmetrical design that gives more clearance towards the RAM slots ensures good compatibility despite of the cooler’s size. The NF-P12 fan can be moved upwards or left off in order to further improve compatibility. | These points suggest good attention to practical details. |
| Excellent component cooling Hanging out at the bottom of the fin-stacks, the NF-P14 fan doesn’t only contribute to the NH-D14’s superb CPU cooling capabilities but also provides massive airflow over surrounding motherboard components and heatsinks, thus ensuring excellent component cooling performance. | Definitely worthwhile for the VRM and chips around the CPU; they run hotter with a hotter CPU. |
| SecuFirm2™ multi-socket mounting Noctua’s enthusiast grade SecuFirm2™ multi-socket mounting provides broad socket compatibility (LGA1366, LGA1156, LGA775, AM2, AM2+ and AM3) and meets the highest demands in safety, contact pressure and ease-of-use. | A secure, tight and easy to use mounting system is essential for any heatsink. |
| NT-H1 thermal compound Noctua’s much-acclaimed NT-H1 is a well proven pro-grade TIM solution that provides minimum thermal resistance, excellent ease-of-use and outstanding reliability. | For continuity, we won’t be using their thermal compound. We’ll take their word for it. |
| Noctua NH-U12P: Specifications (from the product web page) | |
| Socket compatibility | Intel LGA1366, LGA1156, LGA775 & AMD AM2, AM2+, AM3 (backplate required) |
| HxWxD (w/o fan) | 160 x 140 x 130 mm |
| HxWxD (w/ fan) | 160 x 140 x 158 mm |
| Weight | 900g HS only, 1070g w/1 fan, 1240g w/ 2 fans |
| Material | Copper (base and heat-pipes), aluminium (cooling fins), soldered joints & nickel plating |
| Fan compatibility | 140x140x25 & 120x120x25mm |
| Scope of Delivery | 1x NF-P14 premium fan 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 | 12V |
| MTBF | > 150.000 h |
PHYSICAL DETAILS
The NH-D14 retains much of the characteristics features of the
Noctua line: Slightly thicker than usual fins, waved edges on the fins, nickel
plating, and soldering of the fins to the heatpipes. Despites the initial shock
of its size, the overall height of 16cm is no greater than most tower heatsinks
and it should fit without issue in most enthusiast PC cases.
|
|
The heatpipes are centrally positioned and evenly spaced in each
bank of fins. The fan cables are beautifully sleeved, by the way, all the way
to the connectors.
|
|
The fins measure about 0.43mm thick with gaps of approximately
2.3 mm, similar to previous Noctua models. They are fairly thick and the spacing
is moderate compared to other CPU coolers.
|
|
Spring-loaded bolts are prefixed on wings that extend from the base. The bolts
are slight different from most in that they have "female" threads;
it looks like a bolt but acts more like a nut.
|
|
The base is flat and smoothly finished, with only very light machine marks.
Solder is visible around contact points between the heatpipes and the bottom
fins, as well as on the top fins, suggesting every fin is not only press fitted
(like on most heatpipe heatsinks) but also soldered (unlike most).
|
|
INSTALLATION
Noctua split the mounting hardware into three different bags with
AMD parts in one, Intel in another, and a set with parts common to both platforms.
The AMD set requires removal of the standard plastic HS retention bracket that
comes prefitted on most AMD boards, and installation of a metal retention bracket
to which the NH-D14 bolts. The vast majority of AMD boards come equipped with
the motherboard backplate that is required.
The Intel mounting kit is a bit more complex, as it is designed
for compatibility with all three current Intel sockets: 775, 1156 and 1366.
The difference among these three is just the distance between the mounting holes,
which are positioned at the corners of a square.
|
|
The supplied backplate has three marked positions on each wing,
one for each of the Intel socket types. A threaded bolt is fitted to the correct
position on each corner. A rubber cushioning/insulating pad is at each order;
this pad does double duty as a temporary friction fit holder for the bolts during
installation. This is very clever indeed.
|
|
The entire installation process is meticulously thought out, and
trouble-free. It took perhaps five minutes for us to complete, with no concern
at any point about applying too much or not enough pressure on any single point.
An important point here is that the mounting kit supplied with
the NH-D14 is not unique to the flagship heatsink; it is part and parcel of
all the heatsinks for desktop systems that Noctua offers today. The high quality,
security, ease and thorough systems design of Noctua’s approach to heatsink
mounting comes as close to SPCR’s expressed ideal than any other heatsink maker
has come before. From our 2010
HS Test Plaform article:
The mounting mechanism is mentioned because it maintains the all-important
contact between CPU and heatsink. The amount of pressure brought to bear on
the interface also affects cooling. It is also the only real interface between
HS and user. We may say we use a HS, but it’s
not the same way that we use a car, for example. We interact
constantly with a car while using it. User interaction with a HS really happens
only when the HS is installed or uninstalled. If this design aspect is poor
and results in the user having difficulty with installation, or failing to
mount the HS correctly, then poor cooling of the CPU can result. Some mounting
mechanisms are poor, both difficult to install and lacking in precision or
security; others are integrated wonderfully into the heatsink and easy to
use. The mounting mechanism is a critical part of the HS design.
With its latest mounting apparatus, Noctua meets or exceeds all of our expectations
for enthusiast CPU coolers.
|
|
The wire fan fan clips also deserve a closer look, as they are
simple, unusually well-designed and work very well.
|
|
TESTING
We begin with some basic physical measurements.
| Noctua NH-D14: SPCR Measurements | |
| Weight | 940g (heatsink alone) 1080g (w/ 14cm fan) 1100g (w/ 12cm fan) 1240g (w/ both fans) |
| Fin thickness | ~0.43 mm |
| Fin spacing | ~2.33 mm |
| Vertical Clearance (northbridge) | ~53 mm (measured from the PCB to the bottom fin) |
| Overhang (on power supply side of typical case) | negative 11 mm (as mounted with fans blowing toward the motherboard I/O or back panel of case); the fins stayed within the boundary of the motherboard by 11mm. |
A comparison of fin thickness and spacing among various tower
heatsinks is interesting. The NH-D14’s fins fall in the middle, neither tight
nor loose.
| Comparison: Approximate Fin Thickness & Spacing | ||
| Heatsink | Fin Thickness | Fin Spacing |
| Scythe Ninja 2 | 0.39 mm | 3.68 mm |
| Thermalright HR-01 Plus | 0.45 mm | 3.15 mm |
| Noctua NH-U12P | 0.44 mm | 2.63 mm |
| Noctua NH-C12P | 0.47 mm | 2.54 mm |
| Thermolab Baram | 0.44 mm | 2.52 mm |
| Noctua NH-D14 | 0.43 mm | 2.33 mm |
| Prolimatech Megahalems | 0.50 mm | 2.00 mm |
| Xigmatek HDT-S1283 | 0.33 mm | 1.96 mm |
| Scythe Kabuto & Zipang 2 | 0.34 mm | 1.94 mm |
| Scythe Mugen-2 | 0.31 mm | 1.89 mm |
| ZEROtherm Nirvana | 0.43 mm | 1.82 mm |
| Zerotherm Zen | 0.37 mm | 1.80 mm |
| Zalman CNPS10X Extreme | 0.42 mm | 1.50 mm |
| Thermalright Ultra-120 | 0.45 mm | 1.42 mm |
Testing was done on our new
i7-1366 heatsink testing platform. A summary of the test system and
procedure follows.
Key Components in Heatsink Test Platform:
- Intel Core i7-965 Extreme
Nehalem core, LGA1366, 3.2GHz, 45nm, 130W TDP. - Asus
P6T SE ATX motherboard. A LGA1366 X58 chipset board with short solid-state
capacitors around the CPU socket, low profile northbridge and VRM heatsinks,
and mounting holes for both LGA1366 and LGA775 coolers. - Asus
EAH3450 Silent graphics card. - Intel
X25-M 80GB 2.5" solid-state drive. Chosen for silence. - 2GB QiMonda
DDR3 memory. 2 x 1GB DDR3-1066. - Seasonic X-650 SS-650KM
650W ATX power supply. This PSU is semit-passively cooled. At the power levels
of our test platform, its fan does not spin. - Arctic Silver
Lumière: Special fast-curing thermal interface material, designed
specifically for test labs. - Nexus 120 fan (part of our standard testing methodology; used when
possible with heatsinks that fit 120x25mm fans)
The system is silent under the test conditions, except for the CPU cooling
fan(s). The CPU/system is run in two modes:
- Default – using all the stock settings in the BIOS. The AC power
drawn by the system is 165~175W, depending on how well the CPU is cooled.
When the CPU runs hotter, so does the Voltage Regulator Module on the motherboard
which powers the CPU, and it becomes less efficient, causing higher system
power consumption. - Overclocked – to 3.6 GHz, 1.4V Vcore. This is to simulate typical/extreme
overclocking by an enthusiast/gamer. The increased core voltage is is source
of most of the increased power draw. The AC power of the system measures 215~245W,
depending on how how the CPU/VRM is allowed to run.
There is no way for us to accurately determine the power drawn by the CPU.
It is extremely difficult to isolate the CPU’s power because the i7-1366 draws
power from both the AUX12V socket as well as the +5V and +3.3V lines on the
mainboard. We can only guesstimate it to be at least 110W in default mode, and
perhaps ~160W at our overclocked setting.
Normally, our reference fan is used whenever possible, the measured details
of which are shown below.
| Reference Nexus 120mm fan Anechoic chamber measurements | ||
| Voltage | SPL@1m | Speed |
| 12V | 16 dBA | 1100 RPM |
| 9V | 13 dBA | 890 RPM |
| 7V | 12 dBA | 720 RPM |
| 5V | 11 dBA | 530 RPM |
Measurement and Analysis Tools
- Extech 380803 AC power analyzer / data logger for measuring AC system
power. - Custom-built, four-channel variable DC power supply, used to regulate
the fan speed during the test. - PC-based spectrum analyzer:
SpectraPlus with ACO Pacific mic and M-Audio digital
audio interfaces. - Anechoic chamber
with ambient level of 11 dBA or lower - Various other tools for testing fans, as documented in our
standard fan testing methodology. - SpeedFan,
used to monitor the on-chip thermal sensors. The sensors are not calibrated,
so results are not universally applicable. The hottest core reading is used. - Prime95,
used to stress the CPU heavily, generating more heat than most real applications.
8 instances are used to ensure that all 4 cores (with Hyper-threading) are
stressed. - CPU-Z,
used to monitor the CPU speed to determine when overheating occurs; throttling
has been observed to occur at between 95~100°C. - Thermometers to measure the air temperature around the test platform
and near the intake of the heatsink fan.
Noise measurements are made with the fans powered from the lab’s variable DC
power supply while the rest of the system was off to ensure that system noise
did not skew the measurements.
Load testing was accomplished using Prime95 to stress the processor, and the
graph function in SpeedFan was used to ensure that the load temperature is stable
for at least ten minutes. The stock fans were tested at various voltages to
represent a good cross-section of airflow and noise performance.
The ambient conditions during testing were 10~11 dBA and 20~21°C.
FAN TEST RESULTS
|
|
We’ve seen the 12cm NF-P12 before, on the Noctua NH-U12P. The 14cm NF-P14 is
new to us. It is similar in design to the NF-P12, except for size: To maximize
pressure, the nine fan blades have less curvature than typical fans. There is
more of a gap between the blades than in the 12cm model. The "vortex-control
notches," which apparently reduce turbulence and spread the noise generated
from the interaction between the strust and the fan blades over a wider frequency
range, making for a more pleasant sounding fan. The 14cm fan’s frame is fairly
minimal, allowing for the same hole mounting pattern as a standard 12cm fan.
Both fans have good strut geometry; the trailing edges of the blades intersect
with the struts almost perpendicularly, which in theory also helps to minimize
tonal aspect of the turbulence noise. Noctua’s "self-stabilising oil-pressure"
bearing (SSO bearing) is said to surpass ball, sleeve or liquid bearings in
terms of long-term stability and quietness. More details of the fan designs
can be read on Noctua’s web site.
The 14cm fan bearing looks a bit different from the 12cm fan. The bearing end
looks similar to that of the excellent Sanyo Denki San Ace fan employed in the
high end Seasonic X series and MD12 series power supplies. (See
photos on this page.) Noctua has this to say: "Both the impeller hub
and the bearing shell of the NF-P14 are metal reinforced in order to achieve
ultimate manufacturing precision, high durability and vibration-free running."
|
|
The blue marked parameters in the tables below were measured in our lab. Both
fans draw exceptionally low power.
| Brand | Noctua | Power Rating | 0.09A |
| Model Number | NF-P12 | Airflow Rating | 92,3 m³/h |
| Bearing Type | SSO | RPM Rating | 1300 |
| Hub Size | 39.5 mm | Noise Rating | 19.8 dBA |
| Frame Size | 120x120x25 mm | Header Type | 3-pin |
| Weight | 170g | Starting Voltage | 4.8V |
| Brand | Noctua | Power Rating | 0.10A |
| Model Number | NF-P14 | Airflow Rating | 110,3 m³/h |
| Bearing Type | SSO | RPM Rating | 1300 |
| Hub Size | 39.5 mm | Noise Rating | 19.8 dBA |
| Frame Size | 140x140x25 mm | Header Type | 3-pin |
| Weight | 150g | Starting Voltage | 6.5V |
Two low noise adapters are provided, one for each of the fans. The instruction
sheet warns against running both fans off either adapter simultaneously. This
can be done with the splitter cable that’s also provided.
|
|
To get started with the acoustics, noise and speed tests were run on each fan.
Each fan was left mounted on the heatsink (with the other one removed) since
that’s the way they would be used. The noise of both fans running together was
also measured; this is the way the heatsink is meant to be used. Note
the order of the columns in the table below: All the Sound Pressure
Level (SPL, measured in dBA at 1m distance) data is in the middle three columns
for ease of comparison.
| Noctua NH-D14 Cooler: Fan Test Results | ||||||
| NF-P14 fan | SPL, both fans | NF-P12 fan | ||||
| V | RPM | SPL | SPL | RPM | V | |
| 12V | 1230 | 30 | 30 | 24 | 1285 | 12V |
| 11V | 1130 | 28 | 28 | 22 | 1210 | 11V |
| 10V | 1040 | 25 | 26 | 20 | 1100 | 10V |
| 9V | 940 | 22 | 22 | 17 | 1000 | 9V |
| 8V | 840 | 19 | 20 | 16 | 900 | 8V |
| 7V | 720 | 15 | 16 | 13 | 800 | 7V |
| 6.5V* | 660 | 14 | 14.5 | 12 | 680 | 6V |
| 5V | 490 | 12 | 12 | 11 | 570 | 5V |
| * 6.5V is the min. start voltage of NH-P14 fan; to get it to run at 5V, the fan must be started at a higher speed, then the voltage reduced. | ||||||
Our SPL measurements make a mockery of Noctua’s claimed 19.8 dBA at full speed (as well as the ~13 dBA claim with the ULNA). This exaggeration is especially notable when the rest of the specs are so professionally correct.
The 14cm fan is considerably louder than the 12cm model throughout the speed
range. The difference reduces as speed is reduced; by 7V, the noise level from
either fan is soft and gentle enough that the 2 dBA@1m difference is of little
importance.
Each fan has a better qualitative sound when run by itself. Both fans are fairly
smooth, but with a bit of growl and some tonality that becomes more clearly
audible as turbulence noise subsides with reduced voltage. The 14cm fan seemed
a bit smoother in this regard; despite its higher SPL, it somehow made a more
pleasant sound through much of the speed range.
Run together, the sound has more growl and turbulence, and there are some mild
intermodulation effects between the fans, which is almost unavoidable due to
their slightly different speed. At 7V, the overall SPL of the two fans together
matches the reference Nexus 120 fan at 12V. We prefer theNexus for its smoother,
less tonal sound.
At the 6.5V minimum voltage determined by the larger fan, the two fans are
very quiet together. There is considerably less tonality than at 7V.
When the low noise adapters are used as directed by Noctua, the end result
is about 22 dBA@1m, with each fan spinning around 900rpm.
COOLING TEST RESULTS
How likely is the average user going to replace the fans on his new $90~100
CPU heatsink? That’s a relevant question for this review. "Not likely"
seems to be the most likely answer, considering that the stock fans are quieter
than average, and the replacement fans could cost as much as $40 in total.
In that case, at what voltage should the stock fans be set for testing? Keep
in mind that we consider it important to be able to make apples-to-apples comparisons
with other heatsinks we test. Normally, the same reference fan is used at specific
voltages (12, 9, 7, and 5V). This question required some careful thought.
In the end, experimentation with the supplied noise/speed reducing adapters
provided the answer.
- Naturally, a test at full fan voltage/speed is needed: 12V.
- When the blue adapter is used on the 12cm fan, the black adapter on the
14cm fan, and the two fans run together (as per the instructions), the resulting
21 dBA@1m SPL is about same as when a straight ~8.5V is applied to both fans.
Since this would be a common user option, 8.5V looked like a good choice. - If the 14cm fan is run by itself, using the black adapter results in 8.5V
to the fan, and the blue adapter provides 7.3V. We had a third fan
voltage point. (The adapters give slightly higher voltages for the 12cm fan,
8.8V and 7.7V, but to keep the data from overwhelming us, 8.5V and 7.3V were
used with both fans.) - Since 6.5V is the minimum reliable start voltage for the 14cm fan,
this is a natural test point. Users will find a way to get this voltage/speed
if they want to. - The 12cm fan starts at 4.8V, so the usual 5V test point seems a natural,
as it is easy to obtain in a computer.
The SPL values of these various fan settings intersect closely enough with
the SPL obtained from the reference fan or with the stock fans of other fans.
Still, it was decided in the end, to run tests with the reference fan as well.
Many potential buyers of the NH-D14, at least among regular SPCR visitors, will
probably already have a collection of various quiet fans that they may experiment
with. (Most readers know that the Nexus 120 is a rebranded, very slow speed,
sleeve bearing Yate Loon, which is modestly priced just about everywhere.)
There were 10 test runs, with data collected at 4-5 voltage/speed points for
each run. Some runs were repeated because the results did not seem credible.
The time required for these thermal tests and to measure/record all the sonic
permutations extended many, many days.
- With CPU at default settings: A test run with both stock fans, one fan,
then the other fan. - With CPU overclocked/volted: Repeat of the above three sets.
- With CPU at default settings: A test run with one reference Nexus 12cm fan,
then two. - With CPU overclocked/volted: Repeat of the the two sets in #3.
When only one fan was used, it was always placed in the center between the
two fin stacks.
|
|
TEST SET A: Stock fans, default CPU settings
The cooling results with the stock dual-fan configuration are excellent. At
every matching SPL, the stock dual-fan equipped Noctua NH-D14 cools at least
a degree or two (and often, many degrees) better than the closest competitor
equipped with a single fan. As expected with dual push-pull fans, the cooling
performance drops off very little when the fans are undervolted. The difference
between 12V and 6.5V is just 3°C; with either of the fans by itself, the
drops is 6°C or 8°C.
| Noctua NH-D14 w/ Stock Fans, i7-965 default settings | ||||||
| Fan Voltage | NF-P14 only | Both fans | NF-P12 only | |||
| SPL | °C Rise | SPL | °C Rise | SPL | °C Rise | |
| 12V | 30 | 32 | 30 | 31 | 24 | 36 |
| 8.5V | 20 | 34 | 21 | 33 | 17 | 39 |
| 7.3V | 16 | 35 | 17 | 33 | 14 | 41 |
| 6.5V | 14 | 38 | 14.5 | 34 | 12 | 44 |
| 5V† | – | – | 12 | 38 | 11 | 51 |
| *See above text for explanation of chosen fan voltage points °C Rise is the rise above the temperature of the intake air. †Note that the 5V reading for 2 fans can only be achieved by reducing the voltage after the 14cm fan is already spinning at a higher speed. | ||||||
Comparing the dual-fan NH-D14 is a bit difficult because we have not yet tested
any other heatsinks with two fans on the i7 test platform. The only comparative
data set consists of the eight HS tested with a single Nexus 120 fan in our
i7 platform introduction article, which gives at least a starting point.
| °C rise Comparison: i7-965 default settings | ||||
| Heatsink | Nexus 120mm fan voltage / SPL @1m | Rank | ||
| 12V | 9V | 7V | ||
| 16 dBA | 13 dBA | 12 dBA | ||
| Prolimatech Megahalems | 35 | 39 | 42 | #2 |
| Scythe Mugen-2 | 37 | 40 | 43 | #3 |
| Noctua NH-U12P | 38 | 40 | 41 | #3 |
| Thermalright Ultra-120 eXtreme | 38 | 41 | 45 | #4 |
| Zalman CNPS10X Extreme | 39 | 43 | 48 | #6 |
| Thermalright Ultra-120 | 42 | 44 | 49 | #7 |
| Scythe Kabuto | 43 | 48 | 54 | #8 |
| Noctua NH-C12P | 44 | 47 | 54 | #8 |
| Noctua NH-D14 – dual stock Fans | 17 dBA | 14.5 dBA | 38 | #1 |
| 33 | 34 | |||
| Noctua NH-D14 – 14cm fan only | 16 dBA | 14 dBA | n/a | #2 |
| 35 | 38 | |||
| Noctua NH-D14 – 12cm fan only | 17 dBA | 14 dBA | 44 | #5 |
| 39 | 41 | |||
The closest match for the 16 dBA SPL of the reference Nexus 120 at 12V is the
two Noctua fans at 7.3V: 17 dBA. The 33°C rise achieved by the Noctua
NH-D14 at this setting bests the Prolimatech Megahalems by two degrees. But
since there is a 1 dBA advantage for the NH-D14, perhaps the difference should
be considered to be only 1°C. At the lower speed/noise levels, the dual-fan
Noctua’s thermal advantage grows: 4°C over the Megahalems and 3°C
over smaller brethren NH-U12P at 12 dBA.
With the single 14cm fan, the NH-D14 matches the Megahalems at 2nd place, which
suggests that this is the fan that’s really doing much of the cooling. When
only the 12cm fan is used, its performance drops substantially, and its ranking
in the table above falls to 5th place.
TEST SET B: Stock fans, overclocked/volted CPU settings
| Noctua NH-D14 w/ Stock Fans, i7-965 Overclocked/volted | ||||||
| Fan Voltage | NF-P14 only | Both fans | NF-P12 only | |||
| SPL | °C Rise | SPL | °C Rise | SPL | °C Rise | |
| 12V | 30 | 41 | 30 | 41 | 24 | 49 |
| 8.5V | 20 | 48 | 21 | 44 | 17 | 54 |
| 7.3V | 16 | 48 | 17 | 46 | 14 | 60 |
| 6.5V | 14 | 51 | 14.5 | 48 | 12 | 63 |
| 5V† | – | – | 12 | 54 | 11 | FAIL |
| *See above text for explanation of chosen fan voltage points °C Rise is the rise above the temperature of the intake air. †Note that the 5V reading for 2 fans can only be achieved by reducing the voltage after the 14cm fan is already spinning at a higher speed. | ||||||
Even with the i7 overclocked and overvolted, cooling results with the stock
dual-fan configuration remain very good. The lead over the rest of our field
increased with the hotter CPU, with a 3~5°C advatnage over the nearest competitor
at every matching SPL. With just the single 14cm fan, performance also holds
up very well, nearly matching the 2-fan setup, and still besting the rest of
the field. The 12cm fan brought performance considerably lower, as with the
first set of tests.
| °C rise Comparison: i7-965 @ 3.6GHz, 1.40V (Asus P6T SE) | ||||
| Heatsink | Nexus 120mm fan voltage / SPL @1m | i7-965 OC Rank | ||
| 12V | 9V | 7V | ||
| 16 dBA | 13 dBA | 12 dBA | ||
| Prolimatech Megahalems | 50 | 53 | 59 | #3 |
| Thermalright Ultra-120 eXtreme | 51 | 56 | 63 | #4 |
| Noctua NH-U12P | 53 | 56 | 60 | #5 |
| Scythe Mugen-2 | 53 | 56 | 62 | #5 |
| Thermalright Ultra-120 | 56 | 60 | 67 | #7 |
| Zalman CNPS10X Extreme | 60 | 66 | FAIL | #8 |
| Scythe Kabuto | 62 | 70 | FAIL | #9 |
| Noctua NH-C12P | 64 | 73 | FAIL | #10 |
| Noctua NH-D14 – dual stock Fans | 17 dBA | 14.5 dBA | 54 | #1 |
| 46 | 48 | |||
| Noctua NH-D14 – 14cm fan only | 16 dBA | 14 dBA | n/a | #2 |
| 48 | 51 | |||
| Noctua NH-D14 – 12cm fan only | 17 dBA | 14 dBA | 63 | #6 |
| 54 | 60 | |||
TEST SET C: Reference Nexus 120 fans, default CPU settings
To satisfy the dedicated DIY enthusiast’s curiosity, the stock fans were replaced
with two Nexus 120, our point of reference fan. This fan is audibly quieter,
smoother and blends more easily into the background than either of the stock
Noctua fans. It has also often provided better than expected cooling with many
heatsinks, given its slow speed (1100rpm at 12V) and low rated airflow.
|
|
| Noctua NH-D14 w/ Nexus 120 reference fans, i7-965 default settings | |||
| 1 Nexus 120 | 2 Nexus 120 | ||
| Fan/SPL | °C Rise | °C Rise | Fan/SPL |
| 12V/16 | 36 | 31 | 12V/19 |
| 9V/13 | 38 | 33 | 9V/14 |
| 7V/12 | 41 | 36 | 7V/12 |
| °C Rise is the rise above the temperature of the intake air. SPL is cited in dBA@1m | |||
The 3 dBA@1m increase at 12V effected by the second fan follows acoustic theory,
which states that two identical sources of sound will result in a 3 dB increase
in SPL, but unexpectedly, there is only 1 dBA increase at 9V and none at 7V.
A close listening to one vs two Nexus 120 fans at the 9V and 7V settings —
mostly at 6" or closer distance to the fans — led to these conclusions:
- In the anechoic chamber, it is extremely difficult to hear any difference
at 7V, but there is a subtle qualitative difference at 9V, - In an ordinary room, whether any difference is audible depends partly on
proximity effects. If positioned very close to a reflecting surface like a
wall, then the two fans sound a bit louder at 9V. - All the listening confirms that the difference between one and two fans
is at least very subtle. - Inside a typical PC, the difference between one and two Nexus 120 fans at
9V or less is probably not audible for the vast majority of users.
When referenced to SPL, the dual-Nexus fan results with CPU at default settings
are actually better than with the stock Noctua fans. The 31°C rise
best result was also achieved with the stock Noctua fans, but at a noise level
of 30 dBA rather than the 19 dBA of the Nexus fans. This suggests that at lower
noise levels, the Nexus fan produces a bit more airflow that the Noctua fans,
even though the latter have higher maximum airflow capability. It also suggest
that at 12V, the stock Noctua fans produce more airflow than is actually useful;
the CPU temperature might not change even if the fan voltage is dropped to 10V,
for example. The cooling performance drops only 5°C from 12V to 7V.
Even a single Nexus 120 fan does a fine cooling job on the NH-D14, very quietly.
The temperature with one fan at 12V is the same as two fans at 7V, suggesting
that the second fan "doubles" cooling effectiveness.
TEST SET D: Reference Nexus 120 fans, overclocked/volted CPU settings
| Noctua NH-D14 w/ Nexus 120 reference fans, i7-965 OC’d settings | |||
| 1 Nexus 120 | 2 Nexus 120 | ||
| Fan/SPL | °C Rise | °C Rise | Fan/SPL |
| 12V/16 | 45 | 44 | 12V/19 |
| 9V/13 | 52 | 46 | 9V/14 |
| 7V/12 | 57 | 49 | 7V/12 |
| °C Rise is the rise above the temperature of the intake air. SPL is cited in dBA@1m | |||
It’s hard to believe, but the slow Nexus 120 fans manage to provide enough
airflow to this big Noctua heatsink to keep it a top contender at matching SPLs
against all comers even when the i7 is overclocked and overvolted. With these
fans at 12V, the combined SPL is just 19 dBA@1m, yet the 44°C
temperature rise achieved is the same achieved by the stock fans at 21 dBA.
With two fans, the spread between 12V and 7V is just 5°C,
making it easy to choose quieter, slower speed operation — even if the
intake air is a too-hot 40°C, the CPU will remain under the
90°C mark where clockspeed throttling becomes a risk. (Most
PC enthusiasts would be remiss to have 40°C in their case;
temperatures closer to 30°C are easily reachable even in a
quiet, modest airflow case.) An inescapable point: The fact that two Nexus fans
are quieter at the same performance makes one question how good the Noctua fans
are.
MORE COMPARISONS
It’s hard to manage all the data, but here are some snapshots.
A. Stock vs dual Nexus fans
| NH-D14: stock vs. dual-Nexus 120 fans | ||||
| SPL dBA@1m | °C Rise Above Ambient Temperature | |||
| Stock i7-965 | OC’d i7-965 | |||
| Noctua fans | Nexus fans | Noctua fans | Nexus fans | |
| 30 | 31 | – | 41 | – |
| 21 | 33 | – | 44 | – |
| 19 | – | 31 | – | 44 |
| 17 | 33 | – | 46 | – |
| 14 | 34 | 33 | 48 | 46 |
| 12 | 48 | 36 | 54 | 49 |
| All results generated with reference Nexus 120mm fan. | ||||
B. Dual Nexus 120s on Noctua NH-D14 vs Others on single Nexus 120, with
OC’d i7-965
| Noctua NH-D14 w/ 2 Nexus 120 vs. Others on one Nexus 120 | |||||
| Heatsink | SPL – dBA@1m | ||||
| 19 | 16 | 14 | 13 | 12 | |
| °C Rise Above Ambient Temperature | |||||
| Noctua NH-D14 (2x Nexus 120) | 44 | – | 46 | – | 49 |
| Prolimatech Megahalems | – | 50 | – | 53 | 59 |
| Thermalright Ultra-120 eXtreme | – | 51 | – | 56 | 63 |
| Noctua NH-U12P | – | 53 | – | 56 | 60 |
| Scythe Mugen-2 | – | 53 | – | 56 | 62 |
| Thermalright Ultra-120 | – | 56 | – | 60 | 67 |
Of course, the above table begs the question again of how the other heatsinks
would do with two Nexus 120s.
FREQUENCY SPECTRUM SCREENSHOTS
The range of images that can be shown here number in the dozens. Just a handful
are presented, in an attemp to minimize data overload. Note that the fans were
all mounted on the heatsink when these screenshots were captured.
 Both SPL and frequency spectrum are essentially the same with just the Noctua 14cm fan or with both stock fans. |
Two
MP3 SOUND RECORDINGS
The recording begins with the ambient sound of the test room. Please set your
playback volume so that the ambient is almost inaudible, then don’t adjust the
volume control again. For best results, save the sound file to your own PC before
listening.
Noctua
NH-D14 with stock fans The SPLs are approximately 14, 17, 21 and 30
dBA@1m, obtained by setting the voltage to 6.5V, 7.3V, 8.5V (as per ULNA)
and 12V. The recording starts with 7 seconds of the ambient in the room, then
goes through 10 seconds at each of the four voltage settings.Noctua
NH-D14 with stock 14cm fan The SPLs are approximately 13, 16, 20 and
30 dBA@1m, obtained by setting the voltage to 6.5V, 7.3V, 8.5V (as per ULNA)
and 12V. The recording starts with 7 seconds of the ambient in the room, then
goes through 10 seconds at each of the four voltage settings.Noctua
NH-D14 with 2 Nexus 120 fans The SPLs are approximately 12, 14, and
19 dBA@1m, obtained by setting the voltage to 7, 9 and 12V. The recording
starts with 7 seconds of the ambient in the room, then goes through 10 seconds
at each of the four voltage settings.
Comparative Recordings
Scythe
Mugen-2 with stock Scythe fan at one meter
— 8V (11 dBA@1m)
— 9V (12 dBA@1m)
— 10V (15 dBA@1m)
— 12V (24 dBA@1m)Scythe
Mugen-2 with two reference Nexus fans at one meter
— 7V (12 dBA@1m)
— 9V (14 dBA@1m)
— 12V (20 dBA@1m)
FINAL THOUGHTS
The Noctua NH-D14 is a grand statement by a company that specializes only in
high quality CPU coolers and matching fans. The flagship model impresses in
every way:
- High quality manufacturing, including soldering of fins to heatpipes.
- Superbly design mounting hardware for consistent, tight, repeatable installation
in every instance, on both AMD and Intel desktop motherboards. It seems perfectly
safe to mount on any board, despite its high mass. The other HS with comparable
mounting hardware is the Prolimatech Megahalems, but it is only compatible
with Intel socket boards. - Design should fit most systems without issues, despite large size.
- High quality accessories such as ULNA.
- Great attention to details.
- Top cooling results.
The Noctua NH-D14 in its stock form nudges past the Prolimatech Megahelams
for best cooling of a hot, overclocked i7 processor at sub-20 dBA@1m noise level.
The concept of dual-fans on a heatsink in push-pull mode has been explored before
by other manufacturers, but the implementation by Noctua is better than all
the rest thus far. (Thermalright IFX-14 is the only obviously worthy competitor
we haven’t reviewed; a sample has been requested.) The greatest benefit comes
at lower fan speeds, where the usual drop in pressure with a single fan is avoided,
thus maintaining very good cooling at very low noise levels.
Our experiments with the two stock fans suggests that the 14cm fan is doing
most of the cooling work, although the 12cm fan is very helpful at lower speed
with a red-hot CPU. Depite all the features to reduce noise and improve airflow,
the Noctua fans are not exceptionally quiet. Their tonal aspects are sometimes
intrusive, and their ultimate noise-to-airflow ratio is not remarkable.
Far better in this regard are the Nexus 120 fans that were swapped in place
of the stock fans. With the i7 CPU overclocked and overvolted, the two Nexus
fans on the NH-D14 matched the best cooling performance of the stock fans while
making less noise. What makes this result all the more compelling is that the
Nexus 120 is really just a slower version of the Yate Loon sleeve bearing 120mm
fan that’s sold in the US and Canada for $5~7.
Our findings on the combined noise of two Nexus 120 fans will likely spark
another massive discussion in our forums about multiple fans for CPU heatsinks.
It’s highly likely that with a reasonably quiet, smooth fan, the slight increase
in noise of two of them at a low speed (say under 700 RPM) over one is probably
completely inaudible in most PCs for most users. Yet, at low speeds, the cooling
performance will definitely be boosted — assuming your heatsink isn’t already
at the limits of its intrinsic thermal transfer capability (read: not too small).
The question of 2-fan performance on other heatsinks we’ve rated highly with
one fan will have to be addressed, hopefully in the near future, and perhaps
in every new review of heatsinks that accommodate two fans. The cost in additional
time and effort and in review complexity will have to be weighed. In the meanwhile,
users of other highly rated heatsinks such as the Megahalems, the various Thermalrights,
Scythes, Xigmateks, etc. may want to try the simple experiment of adding a second
fan.
Getting back to the NH-D14, Noctua’s flagship model sets the highest standards,
and it is a welcome addition to the range of CPU heatsinks for performance enthusiasts,
especially those who want to experience their power computing silently.
Many of its features really deserve to be applauded; it’s only the relatively
ordinary performance of its fans and the high sticker price which stops it from
getting an Editor’s Choice award.
| Noctua NH-U12P | |
| PROS * Top-notch performance | CONS * Expensive |
Our thanks to Noctua
for the NH-D14 sample.
The Noctua NH-D14 is Recommended by SPCR
* * *
Articles of Related Interest
SPCR’s 2010 CPU Heatsink
Test Platform
Prolimatech Megahalems
Noctua NH-U12P
Thermalright HR-01 Plus
Scythe Mugen-2
Cooler Master Hyper Z600
Intel’s HSF for Core 2 Extreme
Thermalright Ultra-120 extreme
* * *
