DC Load Regulation and AC Ripple
Establishing an accurate load is critical to testing and evaluating a PC power supply. PCPerspective’s power supply test bench can place a precise DC load on the PSU under test. Each power supply is tested under controlled, demanding conditions up to its maximum rated load (at 40ºC). Our current suite of tests includes:
• DC Load Regulation
• AC Ripple and Noise
• Differential Temperature
The SilverStone ST30SF and ST45SF power supplies were evaluated on both features and performance. A full range of equipment was used to test the power supply under controlled load conditions.
• (2) CSI3710A Programmable DC load (+3.3V and +5V outputs)
• (4) CSI3711A Programmable DC load (+12V1, +12V2, +12V3, and +12V4)
• (2) 200W Precision resistor load bank (+12V5 and +12V6)
• Switchable precision resistor load bank (-12V and +5VSB)
• Agilent 34401A digital multimeter (Accuracy ±0.0035% vDC)
• Extech 380803 Power Analyzer (Accuracy ±0.5% of full scale)
• DS1M12 "StingRay" digital oscilloscope (20M S/s with 12 Bit ADC)
• Extech Model 407738 digital sound level meter (Accuracy ±1.5 dB)
The following cables/connectors were used to connect the power supply to the PCPerspective power supply test equipment.
• (1) 20+4 pin ATX
• (1) 8-pin EPS/ATX12V
• (1) 6-pin PCI-E
• (2) SATA
• (2) Molex
• (1) 20+4 pin ATX
• (1) 8-pin EPS/ATX12V
• (2) 6-pin PCI-E
• (2) SATA
• (2) Molex
DC Output Load Regulation
To simulate demanding and maximum loading conditions, the power supplies were connected to the load testers and supplied with a constant 120 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. We have included a second table of expanded tolerances (±1% to ±6%) for reference.
The following tables list the DC voltage results for the SilverStone ST30SF PSU while operating on 120 VAC, 60 Hz.
The ST30SF produced very good voltage regulation on all of the DC outputs with the three main rails staying within ±2% of the recommended ATX guidelines; very good for a mainstream product.
The following tables list the DC voltage results for the SilverStone ST45SF PSU while operating on 120 VAC, 60 Hz.
The ST45SF also produced very good voltage regulation with nearly identical results to the ST30SF; again – very good.
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.
Both power supplies exhibited very good AC ripple and noise suppression with the results staying well below the ATX recommended guidelines, even at full load. As you might expect, the ST45SF power supply exhibited higher values while delivering 50% greater load than the ST30SF.
I disagree with the negative
I disagree with the negative comment on rerating of the operating temp.
Anyone trying to run high power apps in a small enclosure that is typical of a SFX power supply installation “is doing it wrong”. For small enclosures where SFX power supplies are used, the associated motherboard is typically mini-ITX. Anyone trying to run high power apps on a mini-ITX board “is doing it wrong”. I won’t comment on the “fringe users” that attempt to operate high end, heat-producing monster video cards on mini-ITX mobos strapped to a ridiculously overpriced (and maybe even overclocked) CPU in the absolute smallest case they can find just so they can brag about playing the lastest game at the absolute highest resolution possible in a multi-screen setup (with the largets screens possible, of course).
A major problem with “small form factor anything” is moving heat away from the heat-producing elements. The pictures of internals of these power supplies clearly show that these units are “packed” internally. Physical objects block or otherwise restrict airflow making cooling difficult, unless that device is mated to a high volume fan, but then “noise” can become a problem.
Now move this challenge (small size & heat) to a mini-ITX mobo. Pack any “high end” mini-ITX mobo, and there are just a few of those around, with a high power CPU and the problem changes to mounting an adequate CPU heatsink/fan due to component clearances. Since a lot of high power CPUs have extensive I/O options, the end user may even be trading off a lot of capability for small space. Said another way, if a high power CPU has lots of PCIe lanes, and most of those CPUs can power a pair of PCIe x16 Gen2 or even Gen3 slots, how does one access all of those lanes on a mini-ITX mobo? See my point about tradeoffs?
Now combine SFX power supplies and mini-ITX mobos into those small “cube like” cases. The problem of removing heat from the inside of the case becomes a problem of getting enough airflow across all of the heat-producing elements inside the case. If the intent is to build a relatively quiet PC, heat will build up because the high powered fans needed to move the heat away will not be used because they are too noisy. Such a small case could have “mesh” sides all over the place to help with airflow, but then case rigidity (not to mention appearance) could suffer. Again, there are tradeoffs due to size.
Look at the problem this way:
“high power, small size, low noise, low operating temps – pick any 2”
So I fail to see the value in knocking a manufacturer for derating the operating temps of it’s SFX products to something more inline with “common sense” and “reasonable for typical applications”.
Now, the warranty “knock” seems reasonable in view of the general trend in the industry towards longer warranties on power supplies, but again, consider the platform. SFX is a small chassis compared to typical ATX power supplies, so heat buildup may still be an issue here, unless a noisy high airflow fan is used (but then other mechanical engineering issues related to airflow have to be considered), and heat degrades the lifetime of any quality part no matter who makes it.
Also consider the MSRP prices cited in the article itself. At those prices, expecting a longer warranty seems unreasonable nowadays. Longer warranties typically translate to higher product prices because manufacturers will spend more on longer warranty products, both to produce them and to support them through their warranty lifecycles.
So, while I find Lee’s power supplies reviews to be balanced and informative, I think his negative comments in this review are an attempt to “stir up a tempest inside a teapot”.
Talk about stirring up a
Talk about stirring up a tempest inside a teapot! Your rant about people who max out their mini ITX motherboards is over the top. First of all no one is “doing it wrong” except you. Some of us “fringe users” successfully run “high end” mini ITX motherboards inside of those “cube like” cases without problem “even overclocked”. In fact I have been doing it daily for about three years now using the ST45SF power supply without a single component failure. You do it with water cooling and three Cougar Vortex hydro-dynamic bearing fans. High power, small size, low noise, low operating temps – pick all 4! See my point about trade-offs? Oh and I also have 2 8GB GSkill TridentX overclocked DIMMS, 2 4TB WD enterprise 3.5 HDs in RAID-0 (8TB), two Samsung 256GB SSDs, a Swiftech MCP35X water pump, a Black Ice Xtreme 120mm x 38mm radiator, and room for a double slot mini ITX GPU, all running off a high performance ASUS Maximus VI Impact mini ITX motherboard, inside a version 1.0 NCase M1 chassis. At full load for one hour my chassis, mobo, and quad core CPU temps remain under 40C, and at idle are under 32C and totally silent. With modern parts it would be even better. I don’t see your “trade offs due to size” argument, or your “appearance” argument for that matter. You just don’t know how to do it yet.
Ironically, making a case
Ironically, making a case smaller makes it EASIER to get fresh air to components (and hot air out). You don’t even need a battery of case fans lie you do with a 42L ATX Overcompensator: with good design, the fans on individual components are placed right next to the case intakes, so they will pull in their own fresh air directly.
The previous two comments
The previous two comments have small form factor (SFF) builds figured out.
Stop spouting your theories about SFF builds and comment on the power supplies being reviewed.
SFX Joule worrier?
SFX Joule worrier?
As someone who has done
As someone who has done everything you’re moaning about – what’s your issue?
I have a very nice, very quiet HTPC/gaming PC rigged up that’s entirely water-cooled and using one of the older SFX PSUs from Silverstone. Granted it’s very hard to max the PSU even under load with limited component space, but the PSU fan is inaudible or quieter than the already quiet radiator fans.
I think “ease of building” is the main concern – getting everything in neatly was a nightmare, but entirely possible in the FTZ01.
However, at this wattage, most high end users will go for the 500 or 600W versions, so while your gripes miss the mark in general, in this specific case I’d say they’re fair.
I actually found the 40/50degC rating confusing.
WHY is the home unit rated at a lower temperature if the NAS unit would run longer and be more prone to wear out?
Is that the DERATING concern?
Does the derating concern apply or is there some confusion between the NAS and HOME units?
Version 2 is the NAS unit.
Version 3 is the HOME unit.
“450W continuous power output at 50℃ temperature rated for 24/7 operation (V2.0)
450W continuous power output at 40℃ temperature rated for 24/7 operation (V3.0)”
The site next to the FAN PROFILES also says the SAME thing for both units (“perfect for quiet home or office use”) which doesn’t quite make sense since it says above that V3 is quieter and meant for HOME with V2 meant for NAS/IPC.