AC Input
Operating Range:	90-264 VAC .99 power factor
Frequency:	47-63Hz
Current:	10A
Efficiency:	83%
EMI:	FCC-B, CE
DC Output
Output:	+5V @ 30A +12V @ 49A -12V @ 0.8A +3.3V @ 24A +5VSB @ 3A continuous = 610W peak = 670W
Regulation:	3% (+3.3V, +5V, +12V) 5% (-12V)
Ripple:	1% (p-p)
Hold Time:	16ms
PG Delay:	300ms
Safety
OV Protection:	+3.3V, +5V, +12V
OC Protection:	135% OPP
Agency Approval:	UL/ULC/CE/CB/RoHS
Environmental
Temperature:	0° - 40°C
Humidity:	20% - 80% RH
Fan Type:	22 - 55 CFM ball-bearing
Noise:	26-38dB(A)
Miscellaneous
Compatibility:	EPS12V / NVIDIA SLI™ certified
M/B Connectors:	24-pin, 8-pin, 4-pin, dual 6-pin Video
Drive Connectors:	15 (6 SATA, 8 Molex, 1 mini)
MTBF:	100,000 hours
Power Cord:	6' 18AWG (incl.)
Warranty:	3 Years

 

Testing and Performance:

     On the bench we are going to be stressing the Silencer with the following parts:

CPU:	Opteron 170, 2.75GHz, 1.51V
Motherboard:	DFI Lan Party nF4 Ultra-D
GPU:	OCZ 8800GTX, 576MHz Core, 900MHz RAM
RAM:	2x1GB OCZ Platinums, 250MHz, 3-3-2-8, 2.8V
Case:	Danger Den Water Box Plus
Sound:	X-Fi XtremeMusic
Cooling:	CoolerMaster GeminII with 2x120mm Antec TriCool fans
Hard Drives:	1x500GB Seagate Barracuda 2x150GB Raptor X

Voltages:

     First off, we are going to measure voltage stability.  I am using SpeedFan's built in logging capability to capture the voltage readings in realtime, while applying various loads to the system.  I am using Orthos to load both cores on the CPU simultaneously, using Large FFTs.  For the GPU, I used the 3DMark06 test "Deep Freeze", with 8xAA, 16xAF, at 1920x1200.  I randomly applied loads over the course of 30 minutes to see if I could get any significant variance.

     This is the variance from the 12V rail.  The graph shows what is considered acceptable, ie, 12V +/- 5%.  In this case, the rail stayed around 11.65V and varied to just below 11.6 and just above 11.7 volts. That is less than a 1% variance in either direction, which is very good.

     Next up is the 3.3V and 5V rails.  The 3.3V rail hung around 3.25V and pretty much just stayed there.  The 5V rail moved a little bit, but not much, staying solid at 4.81V.

     Vcore was also very stable, only moving from 1.46 to 1.47V.  Next up, we will take a look at power efficiency compared to the Corsair 620HX.

     Power usage was measuered at the plug with a Kill-A-Watt meter.  Idle is defined as being on the desktop, and load was achieved by running Orthos and 3DMark06 at max settings simultaneously.   Here we can see that the Silencer is a more efficient power supply, using 32W less at idle and 5W less under load.  Next up, lets take a look at noise and thermal performance.

     For noise measurements, we used a Radio Shack sound meter, capable of measuring both dBA and dBC weighted measurements.  When testing, the sound meter was placed approximately one inch from the exhaust of the PSU, in order to magnify any effect the PSU noise would have.  In reality, since the user is much furrher away, the effect of PSU noise would be much smaller.  The dBA weighting system emphasizes the 3-6kHz range where the ear is most sensitive.  The dBC scale is more broad and does not attenuate low frequencies nearly as much.  When in idle, both the Silencer and the Corsair perform very similarly.  Under dBC, both have the same value of 64 dB, and the Corsair is slightly quieter in dBA.  When under load however, the story changes a bit.  The silencer becomes 6dBC louder than the Corsair, while still maintaining a 2dB delta in dBA. This would seem to suggest that there is a lot more low frequency noise being generated by Silencer when loaded.

     As for temps, the single 80mm fan actually does a good job of keeping the Silencer cool, with air temperatures consistently 2-3C lower under load than the Corsair unit.