Viewing page 7 of 7 pages.
Previous 1 2 3 4 5 6 7
TESTING & MEASUREMENTS
We did not run any benchmarks on the system. The performance of the 915G and
ICH6-R chipsets would not have been done justice by our conservative
choice of parts. We were more interested in the acoustic properties of this system.
Our testing consisted of running the system under CPUBurn (the heaviest load
for CPUs that we know of) using the Smart Fan setting in BIOS. The Temp Tag
was left at the default 55°C, and then reduced to 50°C to simulate its
performance for a slightly warmer CPU.
Ambient temperature at the time of testing was 20°C. Ambient noise level
was 20 dBA/1m.
|
SB81P
|
|
Temp Tag
|
CPU Temp
|
Case Temp
|
HDD Temp
|
CPU Intake
|
CPU Exhaust
|
System
|
SPL
|
|
Idle
|
37°C
|
35°C
|
37°C
|
450 RPM
|
1250 RPM
|
2000 RPM
|
32 dBA/1m
|
|
55°C
|
58°C
|
37°C
|
39°C
|
1000 RPM
|
1750 RPM
|
2000 RPM
|
34 dBA/1m
|
|
50°C
|
55°C
|
37°C
|
39°C
|
1350 RPM
|
2000 RPM
|
2000 RPM
|
36 dBA/1m
|
Powering up the SB81P runs the fans at full load for a second
or so until the BIOS kicks in and fan speeds are reduced to whatever they are
set to. During this burst (and whenever the fans were run at full speed) noise
was measured at 51 dBA/1m. Presumably, this is the noise level that the system
would be peaking at had we chosen a higher-powered system.
1. Idle
At idle, the SB81P sounds louder than the measured 32 dBA/1m.
The reason is that the twin 60mm exhaust fans produce a decided whine that is
a combination of airflow impedance, aluminum hum and cross-resonance between
the fans. A secondary source of noise was the CPU exhaust fan, and, to a lesser
extent, the CPU intake fan. The power supply did not contribute noticeably to
the noise at this level. Our Barracuda IV also contributed a low rumble that
is audible in the lower frequencies. Even with our conservative system at idle,
we would not call this a quiet system.
Some vibration noise is also evident underneath the fan noise.
We did not take the trouble to locate this noise source, as it was only a minor
irritant compared to the drone of the twin exhaust fans.
2. Load, Temp Tag: 55°C
|
SB81P Heatsink Performance, Temp Tag: 55°C
|
|
CPU Temp
|
Rise from Ambient
|
°C/W tdP
|
°C/W MP
|
SPL
|
|
58°C
|
38°C
|
0.45
|
0.38
|
34 dBA/1m
|
Under the default Smart Fan setting, cooling performance under
CPUBurn was about average for a stock heatsink. However, our results are not
completely reliable because we assumed that the ambient temperature was the
room temperature that the duct was receiving air from. This reflects our best
guess at the air temperature surrounding the CPU.
Although cooling performance was not exceptional,
the temperature did stabilize just 3°C above the Temp Tag. This means the
BIOS fan controller is doing its job: Ensuring that CPU temperature does not
rise above a certain level. It may also simply be a symptom of the relatively
low-powered CPU that we used in our testing, although we are loath to call 84W
tdP low-powered. Still, it is possible that we did not stress the system enough
to reach its limits.
Subjectively, the noise changed considerably from the idle level,
although the only changes in fan speed occurred in the CPU fans. The effect
of the increase was not so much an increase in total volume as an expansion
of the noise across the frequency spectrum. The pitch of the noise is higher
overall, and is centered closer to the musical frequencies to which humans are
most sensitive. The twin system fans remained a major source of noise, although
they did not change in speed.
One thing that was nice about the fan controller is that changes
in fan speed were fairly gradual. Although a small amount of ramping was audible
when we listened for it, it is unlikely that it would be noticed by a casual
user.
3. Load, Temp Tag: 50°C
|
SB81P Heatsink Performance 50°C
|
|
CPU Temp
|
Rise from Ambient
|
°C/W tdP
|
°C/W MP
|
SPL
|
|
55°C
|
35°C
|
0.42
|
0.35
|
36 dBA/1m
|
Even though noise levels were already unacceptable using the 55°C
Temp Tag, we decided to decrease the Temp Tag to 50°C to simulate how a
higher-powered CPU might behave under the Smart Fan setting. Assuming a CPU
that runs 5°C hotter than our sample, this is how we believe the SB81P would
perform.
Cooling performance improved by 3°C over the previous test,
keeping the CPU temperature within 5°C of the Temp Tag. This is a good sign;
it means that a maximum temperature for the CPU can be set with reasonable precision
using the Smart Fan control and an appropriate Temp Tag.
Predictably, noise continued to worsen, with the increase in noise
again coming from the CPU fans. The system fans still did not ramp up, and at
this point were no longer the primary source of noise in the system; the CPU
fans now predominated.
AUDIO RECORDINGS
Because of the distinctive noise that the SB81P makes at boot,
we decided to record a sequence to let you hear exactly how the fans sound
at full speed, and how the fan controller behaves during boot. The last 10 seconds
of the recording reflect the noise level at idle. This recording may be
slightly louder than the measured 32 dBA/1m at idle because the recording contains
hard drive seeks — remember, the computer is booting, and
is therefore accessing the hard drive. The resonance of the twin exhaust fans
is not as obvious in this recording as it was in real life.
NOTE: You may wish to turn the playback volume down before you play the next MP3.
MP3:
SB81P Boot Sequence: Start at 51 dBA/1m, then ramp down to 32 dBA/1m at idle.
A second recording was done with the system under load using the
55°C Temp Tag.
MP3:
SB81P Load: 34 dBA/1m.
Comparables:
MP3: Silverstone LC-11 test system at any load - 23 dBA @ 1 meter SPL
MP3:
Arctic Cooling Silentium T2 with test system (3.5" HDD suspended) at idle - 23 dBA/1m
MP3:
Arctic Cooling Silentium T2 (3.5" HDD suspended) at max load - 34 dBA/1m
MP3: Shuttle SN95G5 SFF system (notebook drive suspended) in idle - 27 dBA/1m
MP3: Shuttle SN95G5 SFF system (notyebook drive suspended) at max load - 30 dBA/1m
|
SPCR MP3s: HOW TO LISTEN & COMPARE
The recordings above were made with a high resolution studio quality
digital recording system. The microphone was 3" from the edge of
the fan frame at a 45° angle, facing the intake side of the fan to
avoid direct wind noise. The ambient noise during all recordings was 20
dBA/1m or lower.
A quick and simple way to use these recordings for valid listening comparisons is to play the quietest recording on only one speaker (or a pair of headphones) and set the volume so it is just barely audible a meter away. You must turn off any special sound effects, and set equalizer / tone controls to neutral or flat. Don't touch the volume setting afterwards, and use the same one speaker when you listen to any of the other files. The end result will be reasonably close to the actual recorded sound levels.
Here is a recording of a very quiet sound that is barely audible from 1 meter away even in a super quiet room.
For full details on how to calibrate your sound system to get the most valid listening comparison, please see the yellow text box entitled Listen to the Fans on page 3 of the article SPCR's Test / Sound Lab: A Short Tour.
|
CONCLUSIONS
The SB81P is sold as a "performance" SFF system. To achieve this
status, it is designed with the latest technology, and supports many bells and
whistles such as PCI Express and multiple
hard drives. Unfortunately, it has become bigger and louder
than its predecessors.
The appeal of a small form factor system is just that: It's small, it doesn't
offend when it's kept on a desk, and it's unintrusive. Putting high performance
parts in a SFF system seems like a bit of a compromise to us. As nice as it
is to be running the latest and greatest hardware, we don't fully understand
why we should buy such a system if it means sacrificing the attributes we appreciate
in a SFF design in the first place.
From a noise perspective, we cannot recommend the SB81P. At 32 dBA/1m, its
noise floor is simply too high to be considered quiet. A large part of the problem
is the use of five fans, three of which are smaller and thinner than the 80
x 25mm design that is the smallest size in which quiet fans are commonly found.
A further problem is the aluminum construction of the case, although this is
a shared failing with almost every SFF system on the market.
If all the loud fans could somehow be quieted, the motherboard does provide a decent
range of options for a quiet system. A full range of undervolting and underclocking
is supported, and the fan controller seems to be fairly effective. It is even
possible that there is enough room across the top of the case to suspend a hard
drive or two.
Much thanks to Shuttle
for providing us the SB81P sample and to Newegg.com
for the Intel 520 CPU loaner.
* * *
Discuss this article in the Silent PC Review Forums.
| Help support this site, buy the Shuttle XPC SB81P from one of our affiliate retailers! |
|
