Silverstone ML05 mini-ITX HTPC case

Cases|Damping
Viewing page 4 of 5 pages. Previous 1 2 3 4 5 Next

TEST RESULTS

Measurement and Analysis Tools

For a case without any fans like the ML05, there can be no baseline noise. It simply doesn't make any noise without components inside. So we go straight to noise, power and temperature measurements for the system installed.

Config 1

At idle load, the system can be easily cooled with the single CPU fan running as low as 1000~1100 RPM, about the equivalent of 4V. The primary noise sources are the hard drive, PSU fan and CPU cooler fan, with some resonance effects from all of them being packed so closely together, and perhaps a touch of vibration effcts adding to the measured SPL. The subjective noise is very low. Assuming placement under or near a large screen TV with viewers seated no closer than ~5' away, the system becomes essentially inaudible as soon as anything is playing over the loudspeakers.

Pushing the CPU using Prime95 is merely a routine exercise, to keep this review in line with other SPCR case reviews. In reality, it's completely unnecessary, as no real user would load a HTPC of this type with such a torture test. In any case, with the CPU fan as jammed under the 3.5" hard drive as it is in our test system, the CPU began throttling about 14 minutes into P95. It cycled every 2 minutes or so: The CPU would cool off to around 70°C after about 90 seconds of throttling, then the CPU would go full load for about 30 seconds before throttling kicked in again. All the other component temperatures also rose, naturally. The hard drive, positioned directly over the CPU, reached 55°C, compared to the SSD, which only got to 46°C.

Playing 1080P video for two hours provoked no change in acoustics and some increases in temperature over the idle state. This is no surprise as total power draw increased by only 10~12W. Still, both the PCH and the HDD ended up running hotter than the CPU, a result of their closed-in position near the CPU.

Video encoding is about the most demanding thing most users will ask of a HTPC. TMPGEnc pushed power demand enough that the CPU fan had to be increased in speed considerably to keep the system from throttling. A reasonable thermal state was reached at 2400 RPM, with the CPU temperature stable at 72°C. All the other component temperatures were elevated but well within safe boundaries. The 24 dBA@1m SPL isn't super quiet by our standards, but it's not bad, and easy enough to take as you can simply walk away while the encoding is being done.

Silverstone ML05
Test Results
State
CPU fan rpm
AC Power
CPU °C
PCH °C
HDD
°C
SSD
°C
SPL
Idle
1100
30W
30
41
38
30
17
HD Video
1100
40~42W
42
50
44
34
17
TMPGEnc
2100
66W
72
63
53
42
24
Prime95
2500
93W
84*
68
55
46
30
*FAIL: CPU throttling, every 2 minutes
SPL
= Sound pressure level in dB, A-weighted, mic positioned 1m at diagonal angle left/front/top of case.
Ambient: 11 dBA, 21°C

It's very clear that this system configuration is one of the worst for the ML05. The 3.5" hard drive blocks the CPU cooler, traps heat between it and the motherboard, and adds its own heat in that area. And even though we've chosen a very quiet drive, its position hanging directly under the biggest opening in the case is hardly helpful for overall noise. All this begs the question of what happens if you don't use the multi-function bracket at all? Now you're limited to the 2.5" drive cage, but there is room for 4 of them in total, which can add up to at least 4TB. Besides, external or network storage is easy to use for a HTPC anyway.

So, out with the WD Red drive, and on with a second round of load testing.

Config 2

Silverstone ML05
w/o Center Bracket or 3.5" HDD
State
CPU fan rpm
AC Power
CPU °C
PCH °C
SSD
°C
SPL
Idle
1100
25W
26
38
28
16
TMPGEnc
1300
59W
60
57
39
19
Prime95
1600
86W
60
55
39
22
SPL = Sound pressure level in dB, A-weighted, mic positioned 1m at diagonal angle left/front/top of case.
Ambient: 11 dBA, 21°C

The minimum noise level for the system hardly changed. But there was a dramatic change under load: In Prime95, the CPU ran a whopping 24°C cooler despite the fan running at just 1600 RPM instead of full tilt, while the noise level dropped 8 dBA to 22 dBA@1m. Curiously, at this load, slowing the CPU fan any more — even stopping it altogether — made no difference to the overall SPL, because of the PSU noise. It appears that inside the case, the PSU fan actually runs considerably faster at this modest load (about 70W DC) than on the open test bench, where it measured 16 dBA@1m. Still, this is not going cost the system points in my book, because P95 is simply not a normal load. The TMPGEnc load is about the highest load this system would see in real use, and there, the CPU fan can be safely run as low as 1300 RPM, for a total SPL of just 19 dBA.



Previous 1 2 3 4 5 Next

Cases|Damping - Article Index
Help support this site, buy from one of our affiliate retailers!
Search: