Thoughts and Performance Results
Thoughts
With each PSU summarised individually, it's time for the pertinent results in graph form so you can easily digest them. What I'd like to point out first, incase is wasn't obvious, is that the vast majority of supplies were able to provide an output power that comfortably exceeds that drawn by modern PC systems available at the time of writing. I cannot emphasise that enough. What holds the majority of capable supplies back from powering a system properly isn't the fact they can't make the power.
It's that the PSU isn't setup by the company that makes it to provide adequate current on certain rails. Factors like voltage ripple have added consequences but the vast majority of issues are down to vendors not tuning the internals of the supply to cope with the demands a modern system will make on a certain voltage rail, usually +12V. The ATX 2.0 spec was created to combat that and the PSU vendors are just getting into their stride when creating supplies with strong +12V abilities.
The vast majority of the ATX 2.0 supplies tested, and ATX 2.0 is what you should be buying these days, will comfortably power a current high-end system without any issues whatsoever. That's great for the consumer and great for the reputable vendors selling high quality PSU designs. Hopefully our roundup has shown you who some of those vendors are and who to avoid.
Supplies are coming to market as you read this with all the connection options and basic supply ability you'll need for systems for quite some time. A handful of the ATX v2.0 supplies on test are also BTX compatible, too, giving you even more future-proofing. When you read some of the excellent reviews out there on the web today that report input power to you, note that the input power you see should be scaled by efficiency to obtain the likely output power the PSU is actually supplying.
When you realise how low that figure actually is, you can see why sites like ours will harp on at you time and time again that it's not the output power rating of the PSU you should be caring about most. It's the connector options that suit your system, quality of output given and the purchase of a quality PSU from a reputable vendor. We cannot emphasise that to you enough. Choose a PSU that fits your budget and has a demonstrable ability to perform well and put the output power rating down your list of considerations. Sorry for the constant emphasis, but the point really does need hammered home to stop you from making poor purchase decisions.
Enough ranting, onto the performance results.
Performance Results
Efficiency
Click for results - 496x734 pixel PNG
Interestingly and one for the conspiracy theorists, the top 4, including two engineering samples given to us by FSP, are all manufactured by FSP Group. The vast majority of supplies have between 70 and 80% efficiency under heavy load however, and 75% is a fine figure to use in quick calculations based on input power, to find output power into the system, if you want a rough guide and can't measure and calculate it yourself.
An efficiency of 80% or more from the FSP Group units is from recent designs that make use of more expensive components that waste less input power as heat, converting more to output power. As time goes by, the other vendors in the group should catch up. 80% efficiency when outputting a true 700 watts is outstanding.
Standby Power
Click for results - 510x735 pixel PNG
The idle power figures will be of interest to those looking to choose a PSU because of financial reasons. Choosing a PSU that's not only efficient but has a low idle/standby power will keep more money in your pocket in the long run. You'll notice that a handful of PSUs are missing from the list. The Tagan TG480-U01, Hiper Type-R Series 480W, OCZ OCZ-520 U12, Akasa AK-P550FF BK and Akasa AK-P650FF BK all had no fixed idle power.
The reading from the 6842A was so erratic that it made it impossible to compute an average figure, with the values jumping around rapidly between nearly 0W and up to ~15W. Without any way to calculate standby power for those, I left them off of the graph. The SSM Technology EVP-5007-00 stands out as a PSU with high efficiency and low standby power.
Look at both graphs together for a better overview.
Voltage Rail Variance
+3.3V Variance - Click for results - 503x739 pixel PNG
+5.0V Variance - Click for results - 505x741 pixel PNG
+12.0V Variance - Click for results - 507x739 pixel PNG
How much the PSU's voltage rails, especially the primary ones, move under load is always something to consider. Related to ripple voltage, stability for your system may be lessened by a PSU whose rail outputs fluctuate too much, causing a certain component in your system to fail due to a voltage too far out of spec. The results in the graphs above are mean averages over time of the rail outputs, measured where possible from the final thirty minute load condition, or over all recorded values if the PSU failed testing for whatever reason.
Tagan, Akasa and OCZ (PowerStream mostly) power supplies stand out here as good performers in that respect.
Temperature Delta
Click for results - 493x737 pixel PNG
Temperature delta is the difference in measured average PSU temperature under load, using the Fluke 52 II, and the recorded average ambient temperature in the lab. These results aren't representative of the temperatures that would result from the PSU being run in an enclosed PC chassis however, and should only be used as a very general guideline. Efficiency plays a large part.
