DC Load Regulation and AC Ripple
DC Output Load Regulation
To simulate real world and maximum loading conditions, the SilverStone SX500-LG PSU was connected to the load testers and supplied with a constant 115 VAC. In this test we are interested in seeing how well a PSU can maintain the various output voltages while operating under different loads.
The ATX12V V2.2 tolerance for voltages states how much each output (rail) is allowed to fluctuate and has tighter tolerances now for the +12V outputs. I have also included a second table of expanded tolerances (±1% to ±6%) for reference.
The following tables list the DC voltage results for the PSU while operating on 115 VAC, 60 Hz.
The SX500-LG PSU produced good load regulation on all of the outputs across the full range of loads. The three primary outputs stayed well within ±3% of their target values, which are what SilverStone claims for this power supply.
AC Ripple and Noise on the DC Outputs
The amount of AC ripple and noise present on the DC outputs was checked using a digital oscilloscope. This AC component may be present in the KHz range where most switching power supplies operate or it may be more prevalent at the 60 Hz line frequency. We adjust the O-scope time base to look for AC ripple at both low and high frequencies. The ATX12V V2.2 specification for DC output noise/ripple is defined in the ATX12V Power Supply Design Guide.
Ideally we would like to see no AC ripple (repetitive) or noise (random) on the DC outputs – the cleaner the better! But in reality there will always be some present. I measured the amplitude of the AC signal (in millivolts, peak-to-peak) to see how well the power supply complied with the ATX standard. The following table lists the ripple/noise results during all of the load tests for the main output voltages of interest.
The SX500-LG power supply exhibited good AC ripple suppression and stayed below 50% of the recommended ATX guidelines, even at full load. However, these values are higher than we are used to seeing with most modern high-quality ATX PSUs.
An important thing to note
An important thing to note about the ‘fanless’ mode: The entirety of the fan control system is composed of a thermistor attached to the internal heatsink that controls fan voltage, and Silverstone using a fan which happened to have a startup voltage that worked well with the voltage produced by the thermistor and desired startup temperature. The problem being that it is an entirely dumb control system, with only the thermal load of the heatsink (or rather, the thermistor strapped to the heatsing with a short length of heat-shrink tubing) providing any hysterisis.
The upshot being that under idle/low-load conditions, the fan will tend to hover just at the startup voltage. Unfortunately, the fan will also make a lot of noise in this regime, with clicks coming from startup/shutdown and buzzing just below startup as energy is dissipated into the rotor sufficiently to produce audible vibration but insufficient to overcome static friction and start the rotor moving.
Sadly, this makes the PSU in practice a lot noisier than it should be. Infuriatingly, without the ‘fanless mode’ the PSU would be really quite silent. Silverstone (via Tony Ou on the HardOCP forums) even tried their own slim 120mm fan, [url=http://hardforum.com/showpost.php?p=1041479391&postcount=596]and rejected it doe to not spinning down[/url], meaning the PSU is semi-fanless at the expense of actual low-noise operation!
There are quite a few other PSU sellers using the same Sirfa platform that Silverstone are using here, and with different fans, so if they avoid the ‘semi-fanless’ mode those should in practice be quieter. Or the SX500-LG could be modified with Silverstone’s own fan if you’re willing to void the warranty (the fan is soldered to the PSU PCB). Option 3 would be to strap a thick 120mm fan to the outside to keep the PSU cool enough that the thermistor never gets hot enough for the fan to approach the startup voltage in the first place.
I’ve replaced the stock fan
I’ve replaced the stock fan in my SX500LG with the Silverstone FW121 that Tony Ou used. It’s not perfect as the fan still has a quiet buzzing/chatter while spinning, but it’s a good improvement over stock.
Also the stock Globefan was not soldered down. It used a 2 pin fan plug+socket, with some sort of glue on top (not engulfing) to secure it.
My next plan is to swap the FW121 for a Scythe Slip Stream SY1225SL12M as it has a very low starting voltage, although this is a full thickness fan that will need to be mounted to the outside.
I recently purchased this
I recently purchased this power supply and it’s installed in my Xigmatek Nebula ITX case with Z97 mobo, 4690k, GTX970 ITX and even while gaming (Shadow of Mordor – high/ultra settings for example) the fan on the psu still doesn’t spin up. I figured out that because my case has negative air pressure, air is being drawn in from the rear of the power supply keeping it cool enough so it never gets to the target temperature for the fan to kick in – making it passively cooled. Fantastic power supply – I’d just rather it didn’t use semi fanless mode as the Nebula case is designed so that the psu aides the removal of warm air. My case, GPS and cpu temps are all fine, I just think they’d be a little lower with the extra exhaust provided by a spinning psu fan.