It’s hard to find an adequate, quiet CPU cooler that will fit in a low profile case. With that in mind, we accumulated a few smallish heatsinks and pitted them against one another and against some of Intel’s stock coolers: Thermolab Micro Silencer and Nano Silencer, Arctic Cooling Alpine 7 GT, and Scythe Big Shuriken.
July 25, 2009 by Lawrence Lee
Stock heatsink/fans that shipped with desktop processors were among the key catalysts that brought SPCR into existence. Their loud, nasty sounding fans
were easily the biggest noise complaint among PC owners, at least until graphics
cards became so powerful that they too required fans. Not only were they bad
to begin with, heat and dust accumulation would inevitably take their toll on
the bearings, and over time, they would get progressively noisier. By comparison,
today’s stock CPU coolers are much improved — they can actually keep the
temperature reasonable (at least at idle) without sounding like jet engines.
Still there are much better, relatively inexpensive alternatives on the market,
and for quiet or silent PC, the stock cooler is typically the first component
thrown aside. However, in some cases, a better, quieter cooler that fits can be difficult to find. Thin is in, interest in low profile cases is on the rise, especially in light of the new popularity of the mini-ITX form factor. By “low profile”,
we generally mean cases that are only tall enough to accommodate low profile (half-height) expansion cards. Typically these cases have no more than 8 cm of heatsink clearance and sometimes substantially less if the enclosure in question is designed with a component (e.g. power supply) overhanging the CPU socket.
Everyone knows that the best CPU coolers are usually the largest, and they have no hope
of fitting into a smaller form factor case. Low profile cases almost always allow just
enough room for the stock heatsink, which becomes the defacto choice
— it’s hard to find a short, quiet CPU cooler. With that in mind, we collected
some smaller heatsinks and pitted them against one another
and against some of Intel’s stock coolers. Represented today are PWM models
from Arctic Cooling, Thermolab, and Scythe. Let’s meet the candidates.
|
Intel Q9550 Stock Cooler (E0 stepping)
Intel’s cooler running 45 nm retail boxed processors ship with a very short
stock heatsink, about 4.5 cm tall. Like the rest of the Core 2 line, the dual
core versions have coolers constructed entirely of aluminum while the quad core
models have copper cores. Our sample comes from a quad core Q9550 purchased in May 2009.
|
|
Fan Measurements | ||
Voltage | Speed | SPL@1m |
12V | 1880 RPM | 26 dBA |
9V | 1290 RPM | 18 dBA |
7.8V | 1110 RPM | 15 dBA |
7.5V | 880 RPM | 14 dBA |
The acoustics of the Q9550’s PWM fan are fairly typical for Intel stock heatsinks.
At higher speeds, it is buzzy with a high-pitched hum. As the speed is decreased
the pitch of the hum lowers and a brushing sound becomes more evident (it sounds
like something soft is pressing against the fan blades). The heatsink features
a fan made by Foxconn which is very sensitive to voltage control under 9V. It
is moderately loud at 12V and quickly becomes quiet below 10V, although its acoustic
character is unpleasant.
Intel Q6600 Stock Aluminum/Copper Cooler (G0 stepping)
Our Q6600 stock cooler sample is about a year old, but has seen very little
service over that time. The Q6600 is a 65 nm chip, and the matching cooler
is a bit larger than the previous one, 62 mm tall and weighing almost a pound (440 grams).
|
|
Fan Measurements | ||
Voltage | Speed | SPL@1m |
12V | 2460 RPM | 31 dBA |
9V | 1810 RPM | 23 dBA |
8V | 1520 RPM | 21 dBA |
7V | 1120 RPM | 16 dBA |
This was the loudest of the three Intel stock heatsinks investigated here, and the
quality of the fan noise was the worst. At 12V it is extremely
loud and turbulent, over 5 dBA higher than the Q9550 cooler. At 9V it is buzzy
and a slight rattle develops. At 8V the rattle causes it to hum and the noise
level approaches 20 dBA which is objectively low, but subjectively, it still
needs to be brought down a notch to be acceptable. At 7V there are two types
of noise, both coming from the motor — an audible hum, and a low-pitched buzzing
sound. Inside a case one meter way it is unlikely to bother most users, but
sitting next to you on a desk in a quiet room, the noise it generates is distinctive.
It should also be noted that we were unable to decrease the fan speed further
than about 1100 RPM — at 6.8V, the fan speed actually begins to increase.
The fan seems to have some type of fail-safe circuitry that prevents it from
running too slowly.
Intel Aluminum Stock Cooler
Our last Intel stock cooler is one of the taller, all-aluminum
types that typically shipped with 65 nm dual core processors. It looks almost
identical to the Q6600 cooler except the fan blades are affixed at a slightly
higher angle and it lacks a copper core. Our aluminum sample measures
1 mm taller (63 mm) and 110 grams lighter (330 grams).
|
|
Fan Measurements | ||
Voltage | Speed | SPL@1m |
12V | 1940 RPM | 25 dBA |
9V | 1420 RPM | 19 dBA |
7V | 1020 RPM | 15 dBA |
6V | 840 RPM | 13 dBA |
The aluminum stock cooler’s acoustics are closer to that of
the Q9550 rather than the Q6600 heatsink. At 12V it produces a breezy hum.
When the speed is decreased a slight rattle presents — this is probably
masked at higher fan speeds by turbulent noise. At 7V there is almost no turbulence,
but the motor’s rattle turns into a more noticeable tick while the fan drones.
At 6V it is almost inaudible. Properly enclosed, we judge it to be quiet at
approximately 8V and below.
Scythe Big Shuriken – US$30
The Scythe Big Shuriken is by far the most expensive and ambitious
cooler in our roundup. It is also the only model to offer mounting on more than
one platform, with AMD and S478 clips included with the standard pushpins The
Big Shuriken is a mini-monster with four heatpipes and relatively massive slim fan
with a diameter of 113 mm. It weighs in at 420 grams and is 57 mm tall.
|
|
|
Fan Measurements | ||
Voltage | Speed | SPL@1m |
12V | 1460 RPM | 23 dBA |
10V | 1230 RPM | 18 dBA |
9V | 1030 RPM | 13~14 dBA |
8V | 820 RPM | 11 dBA |
The Big Shuriken’s slim Slip Stream fan is easily the quietest and best sounding
fan in today’s competition. At 12V it sounds turbulent but smooth and it generates
a small amount of bearing noise. At 10V and below the sound of rushing air dissipates,
while its other characteristics remain.
Arctic Cooling Alpine 7 GT – US$10~$15
The various models in Arctic Cooling’s Alpine line are affordably
priced, offering quiet operation but modest performance — aptly suited to
replace noisy stock coolers. The Alpine 7 GT is the smallest of the
bunch, measuring 64 mm high and weighing in at only 280 grams. It is about
the same height as the small aluminum Intel stock heatsink described earlier, but lighter by a
couple of ounces. The heatsink body has a simple design with a thick base
and broad fins all running parallel in one direction. The fan housing is much
softer than Intel’s and springs when pressure is applied.
|
|
Fan Measurements | ||
Voltage | Speed | SPL@1m |
12V | 2160 RPM | 26 dBA |
9V | 1680 RPM | 20~21 dBA |
7V | 1380 RPM | 15 dBA |
6V | 1160 RPM | 14 dBA |
At 12V, the Alpine 7 GT’s fan is turbulent but smooth, much
smoother than any of the Intel stock units. At 9V the motor begins to rattle.
At 7V the fan develops a gentle, but distinctive hum. At 6V it generates an
fairly innocuous buzzing type sound. We would have no problems running the
fan at 9V or lower. While its acoustics aren’t the best, compared to the other
samples in our roundup, it is excellent.
Thermolab Micro Silencer – US$??
Thermolab offers two different petite-sized heatsinks equipped
with dual heatpipes: the Micro Silencer and Nano Silencer. The larger of
the two, the Micro, sports a large fan with a round frame giving it a higher
fan diameter than the distance between the frame’s corners which is typically
used to rate a fan’s size. The cooler in its entirety measures 58 mm high
and 270 grams.
|
|
Fan Measurements | ||
Voltage | Speed | SPL@1m |
12V | 2650 RPM | 37 dBA |
9V | 1920 RPM | 29 dBA |
7V | 1510 RPM | 20 dBA |
5V | 980 RPM | 15 dBA |
At 12V, the Micro’s fan is extremely loud with its noise profile dominated
by overwhelming turbulence. At 9V, it develops an annoying, constant, low-pitched
tone. At 7V, the fan hums, and considerable bearing noise is plainly audible.
At 5V is exhibits a very dry drone, like the bearing has dried up. The acoustic
qualities are just as unpleasant as the Intel stock coolers.
Thermolab Nano Silencer – US$??
The Nano Silencer is the smallest heatsink in our roundup, weighing in
at 210 grams and measuring 39 mm high. Compared to its larger brother, the
Nano is more square — the fan has a standard 80 mm box frame and is
only 15 mm thick. The top surface of the heatsink has less of a recess compared
to the Micro.
|
|
Fan Measurements | ||
Voltage | Speed | SPL@1m |
12V | 2570 RPM | 31 dBA |
9V | 2140 RPM | 22 dBA |
8V | 1880 RPM | 18 dBA |
7V | 1330 RPM | 15 dBA |
At 12V the fan sounds whiny and turbulent — the fan speed is the same
as the Micro, but the noise level is considerably lower (6 dBA). At 9V it
becomes much smoother, but a high-pitched hum becomes evident. At 8V, the
humming becomes more noticeable as the sound of airflow fades in the background.
At 7V it rattles but is otherwise fairly smooth.
TESTING
Testing was done according to our
unique heatsink testing methodology, and the included fan was profiled
using our standard fan testing
methodology. A quick summary of the components, tools, and procedures
follows below.
Key Components in Heatsink Test Platform
- Intel
Pentium D 950 Presler core, C1 stepping. TDP of 95W; under our test
load, it measures 78W including losses in the VRMs. - Asus P5Q-EM motherboard.
A microATX board with integrated graphics and short solid-state capacitors
around the CPU socket, and a diminutive northbridge heatsink for maximum compatibility. - Intel
X25-M 80GB 2.5″ solid-state drive. - 1GB
of Corsair XMS2 DDR2 memory. 2 x 512MB PC2-8500. - FSP Zen 300W
fanless power supply. - Arctic Silver
Lumière: Special fast-curing thermal interface material, designed
specifically for test labs.
Measurement and Analysis Tools
- Seasonic
Power Angel for measuring AC power at the wall to ensure that the
heat output remains consistent. - 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 sensor. This sensor is not
calibrated, so results are not universally applicable. - CPUBurn
P6, used to stress the CPU heavily, generating more heat than most
real applications. Two instances are used to ensure that both cores are stressed. - Throttlewatch
2.01, used to monitor the throttling feature of the CPU to determine
when overheating occurs.
Load testing was accomplished using CPUBurn to stress the processor, and the
graph function in SpeedFan was used to make sure that the load temperature was
stable for at least ten minutes. The stock fan was tested at various voltages
to represent a good cross-section of its airflow and noise performance.
Physical Measurements
Physical Measurements | |||||
Heatsink | Height | Fan Diameter | Fin Thickness | Fin Separation | Weight |
Alpine 7 GT | 64 mm | 74 mm | 0.44 mm | 2.6 mm | 280 g |
Intel Aluminum | 63 mm | 77 mm | varies | varies | 330 g |
Intel Q6600 | 62 mm | 82 mm | varies | varies | 440 g |
Thermolab Micro Silencer | 58 mm | 85 mm | 0.43 mm | 1.4 mm | 270 g |
Scythe Big Shuriken | 57 mm | 113 mm | 0.32 mm | 1.3 mm | 420 g |
Intel Q9550 | 45 mm | 79 mm | varies | varies | 260 g |
Thermolab Nano Silencer | 39 mm | 76 mm | 0.38 mm | 1.5 mm | 210 g |
TEST RESULTS
Note: All the comparison tables below are aligned to the closest SPL levels.
Intel Stock Cooler: Aluminum vs. Copper Core | |||||
Intel Aluminum HSF | Intel Q6600 HSF | ||||
Fan Voltage | SPL@1m | Thermal Rise | Fan Voltage | SPL@1m | Thermal Rise |
12V | 31 dBA | 24 | |||
12V | 25 dBA | 30°C | 9V | 23 dBA | 29 |
9V | 19 dBA | 37°C | 8V | 21 dBA | 34 |
7V | 14~15 dBA | 46°C | 7V | 16 dBA | 42 |
6V | 13 dBA | 50°C |
The Q6600 heatsink at 9V is quieter and slightly cooler compared to the all-aluminum
heatsink at 12V. This is the only point of comparison where it has a clear advantage.
At the next two lower fan speeds, the Q6600 cooler delivers temperatures 3-4°C
lower while generating an extra 2 dB noise. At best, the copper core design
is only marginally better than its aluminum counterpart.
Mid-sized Battle: Micro Silencer vs. Alpine 7 GT | |||||
Thermolab Micro Silencer | Arctic Alpine 7 GT | ||||
Fan Voltage | SPL@1m | Thermal Rise | Fan Voltage | SPL@1m | Thermal Rise |
12V | 37 dBA | 26°C | |||
9V | 29 dBA | 29°C | 12V | 26 dBA | 38°C |
7V | 20 dBA | 36°C | 9V | 20~21 dBA | 39°C |
5V | 15 dBA | 48°C | 7V | 15 dBA | 46°C |
6V | 14 dBA | 52°C |
The Micro Silencer makes an interesting opponent for the Alpine 7 GT. Both
heatsinks have almost the same height and fin thickness, but the Alpine is taller,
has much wider fin spacing and a smaller fan. At 7V, the Micro generates a similar
level of noise to the Alpine at 9V, but takes a 3°C lead in temperature.
Both coolers can be considered quiet at this level. At 5V the Micro is the same
noise level as the Alpine at 7V, but this time, the Alpine has a 2°C advantage.
As airflow is reduced, the Alpine’s wider fin spacing gives it the edge.
Vertically Challenged: Intel Q9550 HSF vs. Thermolab Nano Silencer | |||||
Intel Q9550 HSF | Thermolab Nano Silencer | ||||
Fan Voltage | SPL@1m | Thermal Rise | Fan Voltage | SPL@1m | Thermal Rise |
12V | 31 dBA | 29°C | |||
12V | 26 dBA | 26°C | 9V | 22 dBA | 37°C |
9V | 18 dBA | 40°C | 8V | 18 dBA | 42°C |
7.8V | 15 dBA | 51°C | 7V | 15 dBA | 49°C |
7.5V | 14 dBA | 55°C+* | |||
*Testing stopped after thermal rise exceeded 55°C. |
Of the two smallest coolers in our roundup, the larger Q9550 stock cooler dominates
the Nano Silencer at 12V, generating 5 dBA less measurable noise while simultaneously
delivering better cooling — 3°C’s worth. The difference narrows at
the 18 dBA level. The Nano does not surpass the Q9550 heatsink until the fan
is dialed down so that it produces 15 dBA of noise. At this point however the
CPU temperatures are so high that we would not recommend using either unit.
TEST RESULTS (Continued)
No Competition: Scythe Big Shuriken vs. Zalman 9300AT | |||||
Scythe Big Shuriken | Zalman 9300AT | ||||
Fan Voltage | SPL@1m | Thermal Rise | Fan Voltage | SPL@1m | Thermal Rise |
9V | 30 dBA | 17°C | |||
12V | 23 dBA | 21°C | 7V | 25 dBA | 18°C |
10V | 18 dBA | 22°C | 5V | 21 dBA | 23°C |
9V | 13~14 dBA | 27°C | |||
8V | 11 dBA | 34°C |
The Scythe Big Shuriken is a mismatch against any of the other coolers in the roundup.
We compared it to a much larger, but underperforming cooler, the
Zalman 9300AT. Despite its considerable height advantage, the Zalman unit at
5V and 21 dBA only just matches the Big Shuriken’s thermal performance at 10V
and 18 dBA. Not only is the Big Shuriken’s measured noise level lower at this comparison
point, its fan sounds a lot smoother as well. This comparison makes it clear that the Big Shuriken is capable of taking
on some of the “standard-size” midrange heatsinks.
Comparison Tables
Performance Comparison at 20 dBA (+/- 2 dB) | ||||
Cooler | SPL | Thermal Rise | Height | Weight |
Scythe Big Shuriken | 18 dBA | 22°C | 57 mm | 420 g |
Intel Q6600 | 21 dBA | 34°C | 62 mm | 440 g |
Thermolab Micro Silencer | 20 dBA | 36°C | 58 mm | 270 g |
Intel Aluminum | 19 dBA | 37°C | 63 mm | 330 g |
Nano Silencer (9V) | 22 dBA | 37°C | 39 mm | 210 g |
Alpine 7 GT | 20~21 dBA | 39°C | 64 mm | 280 g |
Intel Q9550 | 18 dBA | 40°C | 45 mm | 260 g |
Thermolab Nano Silencer (8V) | 18 dBA | 42°C | 39 mm | 210 g |
At the 20 dBA level, none of the other coolers comes close to the Big Shuriken’s
level of performance. The performance is 12°C poorer than
the very best tower heatsink we’ve tested, which is not bad considering its low profile.
There is a large gap before we get to the next best performers: the full-sized
Intel heatsinks and the Thermolab Micro Silencer. These three deliver more or
less the same level of cooling at around 20 dBA. The full-sized Q6600 copper
core cooler is really only more effective with higher airflow.
In last place are the Nano Silencer, Alpine 7 GT, and Q9550 stock cooler. The
Q9550 heatsink performs slightly better than the Nano at 18 dBA, and the Alpine
7 GT subjectively sounds better than the other two, so we give them the edge
over the Nano Silencer. The Nano, with its small heatsink mass and slim fan,
has a hard time keeping up with the competition.
Performance Comparison at 15 dBA (+/- 2 dB) | ||||
Cooler | SPL | Thermal Rise | Height | Weight |
Scythe Big Shuriken | 13~14 dBA | 27°C | 57 mm | 420 g |
Intel Q6600 | 16 dBA | 42°C | 62 mm | 440 g |
Intel Aluminum | 14~15 dBA | 46°C | 63 mm | 330 g |
Alpine 7 GT | 15 dBA | 46°C | 64 mm | 280 g |
Thermolab Micro Silencer | 15 dBA | 48°C | 58 mm | 270 g |
Thermolab Nano Silencer | 15 dBA | 49°C | 39 mm | 210 g |
Intel Q9550 | 15 dBA | 51°C | 45 mm | 260 g |
At the 15 dBA level, we can’t recommend using most of the coolers we
tested today, at least not with a 95W TDP processor. They should be okay with the fans
cranked this low if used with one of Intel’s 65W TDP chips. The Shuriken scores an easy victory, a full 15°C better than the Q6600 cooler.
With low airflow, the Alpine 7 GT shows marked improvement against the rest
of the field. It matches the Intel aluminum stock cooler in performance and
beats the Micro Silencer. Its fin spacing is likely responsible — it is almost
twice as wide as the competition. Its fan also sounds a lot smoother.
The Thermolab Silencers are a step behind the Alpine 7 GT, with the Micro
only performing 1°C better than the Nano. The Q9550 heatsink comes in dead
last.
MP3 SOUND RECORDINGS
These recordings were made with a high
resolution, lab quality, digital recording system inside SPCR’s
own 11 dBA ambient anechoic chamber, then converted to LAME 128kbps
encoded MP3s. We’ve listened long and hard to ensure there is no audible degradation
from the original WAV files to these MP3s. They represent a quick snapshot of
what we heard during the review.
These recordings are intended to give you an idea of how the product sounds
in actual use — one meter is a reasonable typical distance between a computer
or computer component and your ear. The recording contains stretches of ambient
noise that you can use to judge the relative loudness of the subject. Be aware
that very quiet subjects may not be audible — if we couldn’t hear it from
one meter, chances are we couldn’t record it either!
The recording starts with 10 second segments of room ambiance, then the fan
at various levels. For the most realistic results, set the volume so that
the starting ambient level is just barely audible, then don’t change the volume
setting again.
- Arctic
Cooling Alpine 7 GT at one meter
— 6V (14 dBA@1m)
— 7V (15 dBA@1m)
— 9V (20~21 dBA@1m)
— 12V (26 dBA@1m)
- Intel
Aluminum Stock Cooler at one meter
— 6V (13 dBA@1m)
— 7V (14~15 dBA@1m)
— 9V (19 dBA@1m)
— 12V (25 dBA@1m)
- Intel
Q6600 Stock Cooler (copper core) at one meter
— 7V (16 dBA@1m)
— 8V (21 dBA@1m)
— 9V (23 dBA@1m)
— 12V (31 dBA@1m)
- Intel
Q9550 Stock Cooler (low profile, copper core) at one meter
— 7.5V (14 dBA@1m)
— 7.8V (15 dBA@1m)
— 9V (18 dBA@1m)
— 12V (26 dBA@1m)
- Scythe
Big Shuriken at one meter
— 8V (11 dBA@1m)
— 9V (13~14 dBA@1m)
— 10V (18 dBA@1m)
— 12V (23 dBA@1m)
- Thermolab
Micro Silencer at one meter
— 5V (15 dBA@1m)
— 7V (20 dBA@1m)
— 9V (29 dBA@1m)
— 12V (37 dBA@1m)
- Thermolab
Nano Silencer at one meter
— 7V (15 dBA@1m)
— 8V (18 dBA@1m)
— 9V (22 dBA@1m)
— 12V (31 dBA@1m)
FINAL THOUGHTS
Scythe Big Shuriken: The best cooler of the
bunch, in a league of its own. Not only did it claim a substantial performance lead, Big Shuriken easily
has the smoothest, quietest fan. Our only but very serious complaint is the LGA775 stock
mounting mechanism — it is a very poor choice given the tiny space between them and the main fin array above it. Installation is almost a guaranteed
exercise in frustration unless you happen to be gnome.
Arctic Alpine Alpine 7 GT: The Alpine 7 GT is a fairly good choice for
a small LGA775 cooler, especially if it is to replace one of Intel’s low profile
stock cooling units. At higher fan speeds, it is a step below the full-sized
Intel heatsinks, but as the fan speed is decreased, its wide fin spacing gives
it the advantage. Its fan is not the best, but at lower speeds it is fairly
innocuous, especially compared to the competition.
Thermolab Micro Silencer: The Micro Silencer performs about the same
as Intel’s full-sized stock coolers with moderate airflow, but with low airflow
its cooling capability takes a hit, to the point where it struggles just to
beat its smaller brother, the Nano Silencer. Its fan is also poor acoustically,
so it’s not a worthwhile substitute for a full-sized stock heatsink.
Thermolab Nano Silencer: The Nano Silencer’s height made Intel’s Q9550
low-profile copper core cooler its only fair competition in today’s roundup.
Unfortunately, it only outperforms the Q9550 cooler at low fan speeds when cooling
becomes severely compromised. That’s like winning a race to see who can jump
off a cliff first — it doesn’t give it any real bragging rights. Its fan
does sound slightly better, but it is easily drowned out by the sound
of its most useful application: cooking bacon.
To be frank, except for the Big Shuriken and Alpine 7 GT, we wouldn’t use any of
the heatsinks tested today placed in a system on a desk next to us. Their poor
acoustic qualities become noticeable with close proximity, whether you crank
the fans way down or not. At one meter’s distance, they sound adequate at lower
fan speeds — you won’t hear them unless you make an effort, your ears
are sensitive, or you’re in a very quiet room. Also, we would only recommend
their use with 65W processors. The Shuriken is the
only one that can adequately cool a 95W processor quietly.
Our roundup is by no means conclusive; there are many other low profile heatsinks out there. But it seems likely that the niche of a quiet yet effective low profile cooler is underpopulated at this time. If readers have any good candidates for our next low profile heatsink/fan roundup, let us know in the forum discussion thread for this article.
Our thanks to Scythe,
Arctic Cooling,
Thermolab,
and Intel for the
heatsinks used in today’s roundup.
* * *
Articles of Related Interest
Scythe Mugen-2 CPU Cooler
Scythe Katana 3: Same slant, new version
CNPS10X Extreme: Zalman’s Extreme Makeover
ZEROtherm CORE92 Direct-Touch CPU
Cooler
Cooler Master Hyper N520 dual
92mm fan cooler
Prolima Megahalems: A Mega Nehalem
Cooler
* * *