The 2014 El-Cheapo Power Supply Roundup

Comparisons, Conclusions and Fireworks

Graphs

We have another new feature in this year’s roundup – we will now compare how the power supplies tested in this roundup performed on four parameters – 12V Voltage regulation, ripple suppression, efficiency and actual wattage delivered. The power supplies will be sorted from best performer at the top to worst performer at the bottom.

First of all, we will compare the 12V regulation. In this graph, lower numbers are better, as they mean that the voltage stayed more constant, and was closer to the nominal 12V. In this roundup, the Huntkey-built Powercase unit performed the best, with the voltage under by only 1.25% at worst, and dropping by 1.75% throughout the testing, which is actually quite a decent result. The Honli ATX-680 was by far the worst performer in this roundup, with the voltage starting at 12.58V – too close for comfort to the upper limit imposed by ATX specifications.

Next, we will compare the ripple suppression on the 3 main rails. Again, lower numbers are better in this graph. The winner here is the Aywun A1-3000, which comes as no surprise, as it has a complete output filter – consisting of two capacitors and a coil on each rail. The highest numbers recorded from it were 34.2mV on the 12V rail, 12.6mV on the 5V, and 21mV on the 3.3V rail, which gives us an average figure of 22.6mV. The worst performer was the Tsunami 500W unit, with 101.2mV on the 12V rail, 67.2mV on the 5V and 52.2mV on the 3.3V rail. The latter two numbers both violate ATX specifications, and the average figure is 73.53mV. Other units (such as the UltraView, Honli and A-Power) performed worse on individual rails, but their average figures are still lower than those of the Tsunami unit.

The efficiency is another thing we will compare, as this is where cheap and nasty power supplies really fall on their face, since many of them use very dated and inefficient designs. In this graph, higher numbers are better. The best performer here was the Auriga MPT-301, which was the only unit to exceed 81% at any time in the testing. The worst performing unit was the A-Power 680W, which never even made it to 75% efficiency, making it one of the most inefficient units I’ve ever tested.

The final comparison will be what percentage of its rating the power supply was able to deliver with the voltages and ripple within allowable limits. Again, the Auriga unit comes out on top here, being the only unit which was honestly rated. The A-Power was again the worst performer, as it could not deliver a single watt with its ripple below the upper limit on the 5V rail.

Conclusion

So, unlike in previous years, we did actually find ourselves an honestly rated unit from a no-name brand. However, that doesn’t change the conclusion. We still can’t recommend a power supply like the Auriga unit tested here due to its poor voltage regulation, questionable component selection, and dated 5V heavy design – problems that it shares with all other cheap generic units. Additionally, the new Minimal Energy Performance Standards laws in Australia mean that it is now illegal to sell a new PC with any of the power supplies tested in this roundup, as all new PCs must now have 80 Plus Silver or higher certified power supplies in order to comply. If you are on a tight budget for your new PC build, an In Win Power Man unit such as the IP-S400GQ3-2, or an OEM Seasonic unit, such as the SSP-350GT, remains our recommendation.

The Fireworks

Discuss this review in our forum here.

< Previous Page

Pages: 1 2 3 4 5 6 7 8 9 10 11