Prices and Final Thoughts

This roundup shows that many power supplies within this wattage range are some of the best overall choices; users can get very good quality for a reasonable price. Most users don't need even 650W, but it's more difficult to find good quality PSUs in the 300-500W range, and here at least there's plenty of room for future expansion - most SLI/CF configurations will work great on these 650W units. All of the tested units also provide enough connectors to support such setups. While all of these power supplies are good, a few models stand out in certain areas.

Hiper and Xigmatek both provide four PEG connectors, which is enough to support all but the highest-end SLI/CrossFire setups. Running dual 8800 GTX/Ultra or HD 2900 XT cards won't leave much room for other areas, so if you're looking to run an overclocked quad-core CPU along with four or more hard drives you would almost certainly want to move beyond 650W PSUs. However, for stock performance the Hiper and Xigmatek can handle just about any CPU/GPU configuration currently on the market.

The efficiency rating of all the power supplies is also very good, peaking at 85% or more. be quiet! tops the list with up to 87% efficiency and Seasonic comes next with 86% - impressive considering Seasonic has been selling this unit for over a year. All the other power supplies are only fractionally behind, however, offering 84% efficiency up to 500W load.

In terms of noise levels, the be quiet! unit actually manages to live up to its name. At a maximum of only 21dB(A), this power supply is not audible and even with full load and you need to get very close to the fan to hear it. The fan speed is about 500RPM, which is definitely slow and helps to keep noise down, but it also results in less air moving through the power supply. Hiper and Seasonic also perform well with only 20dB(A) as long as they stay under 500W of load. The others are all above 25dB(A) from 500W onwards, but they stay below 30dB(A) and are still quiet enough for all but the most demanding Silent PC users.

The build quality is excellent with all the power supplies we looked at today. We didn't find any flaws during operation and none of the units went up in flames. (Ed: Drat!) When we consider the fact that these are all moderately expensive PSUs and three are manufactured by the same ODM, however, this is not a surprising result. Many users would be well served by a power supply from this wattage range, provided you don't skimp on the price. You get decent quality capacitors and other components, and the results speak for themselves. The prices are higher than many other lower quality products, but when you consider how much money is spent on CPUs, GPUs, RAM, and other areas, it hardly makes sense to save a few dollars on the component that makes everything actually function.


Prices


Prices

In Europe, the prices are all quite similar, differing only by a few Euros. It's strange to see the Thermaltake Toughpower Qfan for around 100 EUR - the cheapest in Europe - but it's $160 in the US, which is the most expensive option. While none of the power supplies are bad, a few look to be better than the rest. The Hiper and Xigmatek offerings provide a great balance of price, features, and performance, whether in Europe or the US. Thermaltake is also worth a shot in Europe, given it's 10 EUR cheaper in some areas, and Antec warrants a look in the US at $120. The be quiet! Dark Power PRO ends up as the most expensive unit in the US and Europe, but at the same time there are users that will pay the money for silence.

Our overall recommendation remains with the Hiper HPU-5B680 and Xigmatek NRP-MC651. Both performed very well throughout testing, no doubt helped by the high quality components such as Hitachi and Nippon Chemi-Con capacitors. Choosing between these two options in the US will come down to personal preference - slightly lower noise (Hiper) or a better price (Xigmatek).

Comparison: Efficiency and Acoustics
Comments Locked

23 Comments

View All Comments

  • larson0699 - Monday, August 4, 2008 - link

    You mentioned that the label on the Thermaltake unit doesn't state maximum combined load of the lower-volt rails.

    But look, it does. 180 Watts.
  • BRDiger - Tuesday, July 29, 2008 - link

    I kind a wondering because comparing the efficiency level of the Be Quiet! to the testing datasheet on the 80plus.org site it just reaches about 80% at 100% load unlike your results with nearly 84%. Comparing it to the Enermax Modu 82+ review here, it even exeeds its efficiency
    Just wondering and it´s not supposed to be an offence...
  • BRDiger - Tuesday, July 29, 2008 - link

    Well, after reading another review it seems that 80plus was actually wrong and the be quiet!s performance exeeds the enermax..
  • strikeback03 - Friday, February 22, 2008 - link

    Thermal and fluids was not my best subject in school, but don't fans usually lose air off the tips of the blades, not suck it in? That is my experience with the Zerotherm Nirvana, which seems to make the same claim of sucking air in with its open fan.
  • mo3 - Friday, February 22, 2008 - link

    The new Enermax Modular 82+ & Pro 82+ can peak over 700Watts with out a problem; also they have the ATX12v V2.3 and reaching 88% efficiency.
    Modular 82+ has 2x 12 pin connection embedded for up and coming graphics cards and greater stability.

    Just to give the readers a better alternative! :)

    Check out more info at: www.enermax.co.uk
  • crimson117 - Friday, February 22, 2008 - link

    Okay... PSU review.. new 9600 GT review... I think anandtech is trying to tell me I need a new computer!
  • NINaudio - Thursday, February 21, 2008 - link

    Hi, you mention that the be quiet! PSU is quiet but also runs warmer than the others. Where are the temperature charts that we've seen in previous PSU reviews? It would be good to see the temperatures that you sacrifice for a quieter PSU. Is it a degree, 5 degrees, ten degrees, more?
  • tynopik - Thursday, February 21, 2008 - link

    1. include a real stress test like how the units handle rapidly changing input voltage

    say load them up to 80% capacity and then connect them to a variac and dip the voltage to ~90V 3 times in quick succession to simulate a brownout or poor electrical conditions and check that
    a) it doesn't blow up
    b) it maintains quality output while the input if fluctuating

    2. include cheapo units for comparision just for the reminder of what a poor quality unit is

    3. i really love your connector length charts, far more useful than the usual 'stetch a jumble of cables out and lay a ruler next to it' approach. i just wish you would do the same for the sata/peripheral connectors

    4. i'm really like the way you combined all the acoustic and efficiency charts together instead of having a bunch of individual charts
  • Christoph Katzer - Thursday, February 21, 2008 - link

    I just assembled a nice system do to those tests. Once it's up and running I will include it. Will also make a comparison of all cables from now on, thanks.

    The problem with cheap units is that they are mostly for a specific market. I would need to buy thise units myself since no company would ever send them over. Since I am not in the US I will only have the stuff from Europe which is not even available in the US. But let us see how we could change that in the future...
  • jtleon - Thursday, February 21, 2008 - link

    Chris,

    You state that at least 100W must be drawn before these units achieve high efficiency. Can you relate this requirement to most normal PC tasks - i.e. browsing the web (reading Anandtech reports of course), answering email, other office based activities (the majority of PC users on the planet), rather than those hardcore gamers that are gaming around the clock (and clearly are independently wealthy!).

    It would appear that these PSU's are not going to efficiently reduce the carbon footprint of the majority of PC users on this planet.

    Am I wrong?

    Regards,
    jtleon

Log in

Don't have an account? Sign up now