Viewing page 3 of 3 pages.
Previous 1 2 3
TESTING
Testing Platform
The SS-300FS was installed in one of my main systems, which I would describe
as a stable, very quiet, low airflow PC running Windows 98SE, fully updated:
| Case |
Landmark ATX-202 18" tower |
| CPU |
Pentium 4 - 1.6A (overclocked to 2.0 GHz) |
| Motherboard |
Gigabyte GA-81RXP |
| RAM |
512 MB 2100 PC DDRRAM |
| Video Card |
Matrox G400 Max (dual mode, driving two 19" monitors) |
| Hard Drives |
Seagate Barracuda IV - 20 G |
| |
Seagate Barracuda IV - 40 G |
| |
Both drives in cage at bottom of case, behind inflow case
fan |
| Floppy Drive |
Generic |
| DVD Drive |
Toshiba SD-M1502 |
| CD-Writer |
Creative RW121032E |
| Network Card |
Intel Pro/100VE - Built-in to motherboard |
| Sound Card |
Creative SB PCI128 - Built-in to motherboard |
| Fans |
1 Panaflo 80mm "L" @5V over stock Intel heatsink |
| |
1 Panaflo 80mm "L" @5V over video card / NB heatsink |
| |
1 Panaflo 80mm "L" @4V lower front case fan |
Test Instrumentation and Environmental Conditions
|
CPU temperature
|
Motherboard Monitor 5 reading CPU diode |
|
PSU temperature
|
Veriteq Spectrum 1000 with probe lodged in PSU
heatsink |
|
Fan voltage
|
Heath / Zenith SM-2320 Multimeter across fan terminals |
|
System power
|
Calculated with Super
Simple Power Meter |
|
Noise*
|
Heath AD-1308 Real Time Spectrum Analyzer |
| Ambient temp |
23° C |
| Ambient noise |
32 dBA ("A" weighted, approximate) |
Motherboard
Monitor 5 needs no introduction for most PC enthusiasts. It enables monitoring
of temperatures and voltages off motherboard sensors. With the diode being read,
there is little question of inaccuracy here; it is usually accurate within 1°
C. The Veriteq Spectrum
1000 will be familiar to anyone who has read my PSU and hard drive articles
on this site: it is a highly accurate data logger that samples temperatures
via its probe. The sampling frequency can be varied from every 10 seconds to
once a day. The Heath is an ordinary multimeter that has proven to be fairly
accurate.
The Heath AD-1308 is a portable half-octave Real Time Spectrum Analyzer with
SPL meter functions. Below 40 dBA, its accuracy is limited to 3 dB increments,
down to 23 dBA. Some 15 years old, this LED-based unit has long since been
displaced by digital devices with better interfaces to PCs. However, it is one
I have access to and doesn't cost hundreds of dollars to buy. (Thanks, Tommy!)
The "A" weighting was used, as recommended by numerous acousticians.
*NOTE on Noise Measurements: The mic was positioned almost touching the back case panel, about an inch to the side of the PSU fan exhaust to avoid fan turbulence in the microphone itself. The dBA measurements obtained here cannot be compared to any other measurements due to the lack of adherence to a repeatable standard and an uncontrolled reflective environment. (Sorry,
I tried, but I will not have access to that anechoic chamber till September.)
No effort was made to change acoustics in my small office, which measures 12 by 10 feet, with an 8 foot ceiling. The PC sits on the floor, under table that supports the monitors.
At Startup
Subjectively, the PSU is quiet, but audible. Considerably more audible than
my modified power supplies, which run Panaflo 80 mm fans at 5V or less, and
are essentially inaudible most of the time (depending on ambient noise). It's
an obviously unfair comparison, because the Panaflo is rated at 21 dBA at full
tilt, while the ADDA in the Seasonic is rated 13 dB higher! Still, that is my
ideal reference for a fanned PSU.
The following measurements were taken after Windows 98SE was running for 2
minutes, with no programs open:
| CPU temperature |
40° C |
| PSU temperature |
35° C |
| Fan voltage |
4.31 V |
| System power |
72W |
| Noise* |
41 dBA (A) |
With Hard Use
Time to see what happens when the system is pushed. I opened up my most demanding
applications: Photoshop 6.5 and Adobe Framemaker 6.0, and opened large files
in both. For good measure, I turned on my mail program and opened MS Internet
Explorer to the default home. I then turned on the only real game on the system,
Command & Conquer: Red Alert 2 and proceeded to play it in
the main 19" monitor. The CPU temperature climbed steadily and quickly to 55-56° C within a few minutes, but then remained unchanged after that. The following
readings were taken 15 minutes after the game was started.
| CPU temperature |
55° C |
| PSU temperature |
40° C |
| Fan voltage |
4.32 V |
| System power |
~80W to 98W |
| Noise* |
39-40 dBA (A) |
As you can see, the voltage to the PSU fan did not change. The 0.01V change
recorded is well within error tolerances and does not represent a real change.
I certainly did not hear any difference in the noise between startup and at
this point.
I tried a slightly different tack: writing this article in Macromedia Dreamweaver
4 while leaving all the above programs on -- including the game. Now I began
to see and hear some changes. I continued pushing my system for another 15 minutes.
Here are the results:
| CPU temperature |
57° C |
| PSU temperature |
42° C |
| Fan voltage |
4.38 V |
| System power |
96-98W |
| Noise* |
40-41 dBA (A) |
The fan was only very slightly but definitely louder. The change hardly registered
at all on the sound meter; I believe it amounted to a maximum of 2 dB, perhaps
as little as 1 dB. After another 15 minutes, I established that the temperature
could not be pushed beyond these levels. Fan noise remained unchanged.
With CPU Stability Test
CPU Stability Test by Jouni Vuorio
is a useful tool to stress systems. I set it to CPU warming only after shutting
down all the other programs and allowing it temperatures to go back down to
those at startup. Here are the results over 45 minutes.
System power is not provided as it stayed at a constant 120W.
| Time (min) |
0
|
5
|
10
|
15
|
20
|
25
|
30
|
35
|
40
|
45
|
| CPU (° C) |
40
|
60
|
63
|
64
|
64
|
65
|
65
|
65
|
65
|
65
|
| PSU (° C) |
35
|
39
|
42
|
43.5
|
43.7
|
43.9
|
44.1
|
44.8
|
45
|
45.2
|
| Fan (VDC) |
4.32
|
4.32
|
4.38
|
4.51
|
4.62
|
4.65
|
4.67
|
4.79
|
4.82
|
4.84
|
| Noise dBA* (A) |
39
|
39
|
40
|
41
|
41
|
41
|
42
|
42
|
43
|
43
|
The measured noise difference between start and finish was about 4 decibels.
I have to say this is definitely an approximation, as environmental conditions
were not ideal. What I heard jibes with that number -- the fan did get louder,
and this difference was substantially louder than before. I personally would
not be satisfied with the sound level of the PC at the end of this test. However,
running torture programs is not something the usual PC user is likely to do
often, and the noise that I heard was substantially less than I've experienced
in the past with other thermistor-controlled fan PSUs.
It is interesting to note that 41-42° C appears to be the heatsink temperature
at which the fan voltage begins to rise above default. After the CPU hit 65°
C, the PSU temperatures continued to rise while the CPU temperature remained
unchanged. I believe this has to do with the buildup of hot air from the CPU
/ heatsink which can only exit through the PSU. Left long enough,
what would likely happen is that the temperature of the PSU would increase high
enough to a point where the fan would also be increased high enough to stabilize
the PSU temp, and then an equilibrium would be reached. As I feel any kind of
prolonged torture testing is an unrealistic load, I did not bother to see when
this would happen in the test system; the result would be very system dependent anyway.
CONCLUSION
The Seasonic SS-300FS is an impressively quiet power supply that remains at
the same quiet level throughout a wide range of conditions in a system that
while not heavily loaded, is somewhat more full featured than average systems.
When the fan did speed up, it did so only under extreme conditions of use or
during CPU stress testing, and the degree of noise increase witnessed and
measured was modest. The fan voltage never actually reached anywhere close to
12V; it remained under 5V throughout the testing.
There was never any question of stability being an issue: I experienced not
one crash or other computer misbehavior during the several days this PSU has
been in my system.
Compared to ALL other unmodified power supply units I have used, the Seasonic
SS-300FS is about the quietest. These other units include 2 models of Enermax,
SH, Zalman and many generic power supplies. Despite the slightly quieter startup noise level of the Zalman ST-300BLP, that unit's rapid speedup makes it significantly more intrusive than the Seasonic, overall. As mentioned
before, this is the same behavior observed in the Enermax and SH models, as
described in detail in my previous PSU articles.
The PC Power & Cooling 275W model I have on hand may come closest among the unmodified competitors I have heard. It does not sport a thermistor, thereby avoiding
the distracting annoyance of a variable speed fan, and the fan is rated for
20 dBA in free air, 34 dBA in the PSU. I have not had the opportunity to examine
this unit carefully. A review will be forthcoming in the near future. Other models that looks promising include the Fortron available from PCSilent.de and the Q-Technology models from QuietPC.com.
Seasonic's S2FC fan controller circuit certainly is a big step in the right
direction for quieter power supply design, an innovative and intelligent application of existing technology. The only improvements we can suggest is a quieter fan and a proper wire fan grill.
|
PROS
|
CONS
|
|
|
- Not as quiet as my own modded PSUs
|
|
|
- No track record in North America
|
|
|
|
|
|
|
|
|
|
Conversations with Seasonic
When asked why a quieter fan like the 24 CFM Panaflo fan (21 dBA) was not chosen
for this unit, Seasonic expressed the need to ensure more than adequate airflow
at full rated power for safety reasons. Various agency approvals required for
PC power supplies dictates certain design criteria. The current fan does deliver
more airflow than is ultimately necessary, which Seasonic says is one way to
ensure longevity and reliability, but they are experimenting with quieter fans
to reduce the noise further. Considering that under a real load, the controller
never fed the fan even 5V, I'd say it's a pretty conservatively rated design that would probably be perfectly safe with a quieter fan even with airflow.
I cannot really imagine how the full 300W capacity could be drawn by any desktop PC.
A promising tidbit mentioned was Seasonic's current effort to increase efficiency at
full power from the current ~65% to approximately 75% on all the ATX models before the end of the year.
This would be a most welcome benefit. If the efficiency can be maintained at
lower power levels, the internal heat generated at 100W power delivery would
drop for 53W down to 33W. That is a most significant reduction in heat that
would lead to a dramatic reduction in cooling needs. We all know what means:
a quieter fan.
I am grateful to Seasonic for the review samples, their patience and their responsive assistance. Thank you.
As noted earlier, Seasonic is only just beginning to establish their presence
in North America. Currently, the following resellers have been established.
They do have a full range of models.
Royal
- The Computer Superstore
Case Depot
Silicon Acoustics
PS -- Look here for a review addendum on their 400W model soon!
* * * * *
Discuss this article in our Forums.
| Help support this site, buy from one of our affiliate retailers! |
|
