Voltage Regulation

+3.3V regulation
Load Voltage
10% 3.33V (+0.91%)
20% 3.30V (+0%)
50% 3.27V (-0.91%)
80% 3.25V (-1.51%)
100% 3.21V (-2.73%)
110% 3.18V (-3.64%)

+3.3V is often the weakest rail with ATX-spec loads. A modern system doesn’t need many amps from this rail, but a power supply shouldn’t have problems reaching the advertised performance. In that sense, –3.64% is not the best result, but still it's still not a problem. This rail is very stable from 10-50%, after which the drop becomes larger. The load was 1.58A at 10% and 17.36A at 110%, but modern systems aren't likely to exceed ~50% load on this rail.

+5V regulation
Load Voltage
10% 5.00V (+0%)
20% 4.99V (-0.2%)
50% 4.97V (-0.6%)
80% 4.94V (-1.2%)
100% 4.92V (-1.6%)
110% 4.90V (-2%)

+5V starts at an optimal level with 5.00V. Usually we like this result to start higher, because there is often a large voltage drop under higher loads. There's not much room for a drop, but as you can see, it is still very stable. Even at 10% overload (-2%) all measurements are within the ATX specifications. The load was 1.89A on 10% and 20.83A on 110%.

+12V regulation
Load Voltage (best) Voltage (weakest)
10% 12.17V (+1.42%) 12.13V (+1.08%)
20% 12.21V (+1.75%) 12.15V (+1.25%)
50% 12.18V (+1.5%) 12.12V (+1%)
80% 12.12V (+1%) 12.07V (+0.58%)
100% 12.07V (+0.58%) 12.03V (+0.25%)
110% 12.04V (+0.33%) 12.00V (+0%)

+12V shows an untypical regulation and increases at 20% load with up to 12.21V. However, the voltage is always over 12.00V. +12V5 was always the weakest rail here but still better than most of the results we have measured in previous tests. The load was 8.65A on 10% and 94.95A on 110% (for all rails combined). By the way +12V has an OCP of 50A at all +12V rails, which is quite high.

With the OC version you can adjust the rail up to 12.6V, for those setups where one cable has to transfer very high currents with their accompanying high voltage drops. Most users will have no need of this feature, as the internal regulation is very good.

+5VSB regulation
Load Voltage
10% 5.10V (+2%)
20% 5.09V (+1.8%)
50% 5.05V (+1%)
80% 5.01V (+0.2%)
100% 4.97V (-0.6%)
110% 4.95V (-1%)

The well dimensioned +5V Standby rail starts higher than the other +5V rail. The load was 0.29A on 10% and 3.17A on 110%.

Interior and Topology Efficiency and PFC
Comments Locked

29 Comments

View All Comments

  • bobbozzo - Friday, July 9, 2010 - link

    Most houses in the US have 20A circuits, but 15A outlets.
  • Klinky1984 - Saturday, July 10, 2010 - link

    A lot of people buying this probably don't need 1200 watts. There is some prestige to having a 1.2kW PSU, but what kind of system is really going to stress this power supply? Maybe if you got one of those 7x PCIe x16 motherboards and created a GPU distributed computing number cruncher, perhaps. But I'd think a more typical tri-SLI setup would have problems pushing 50% usage on this thing.
  • Klinky1984 - Saturday, July 10, 2010 - link

    Well, perhaps I underestimated the power-fatty Fermi is.
  • cactusdog - Thursday, July 8, 2010 - link

    This doesnt appeal, there are much better options out there with truly silent 140mm fans. Most new high end PSUs are going back to a single rail too.
  • HOOfan 1 - Thursday, July 8, 2010 - link

    >most new high end PSUs are going back to single rail

    I wouldn't say that at all....some companies want single rail, but I think MOST is a serious exaggeration.

    Why does it matter anyway?
  • Stuka87 - Thursday, July 8, 2010 - link

    If you buy a cheap supply, its going to be cheap. Regardless of the brand. I have a mid-range Antec and I have been very happy with it. Its quiet, and it weighs about 3x more than the cheaper PSU it replaced.
  • doctormonroe - Thursday, July 8, 2010 - link

    I'm not a big fan of having less than a 120mm fan, but the reviewer covered it when he said that you'd hear the rest of the system before a 1200W PSU under load.

    I'm glad that you've stopped using the charts that were in previous articles (as they were not easy to read and comprehend), but good charts are much better than tables, so hopefully soon you'll figure out a good chart to use.
  • HOOfan 1 - Thursday, July 8, 2010 - link

    Yeah...that happens. It happens with anything...especially with electronics.

    What model was it? If it was REALLY one of their halo products, I am not sure how you came to the conclusion that it was built cheaply.

    Of course you have some people who think the PSU is such a simple design and that there is no excuse for a well built PSU to fail. People like that just make me roll my eyes...and usually ignore them from then on.
  • Martimus - Thursday, July 8, 2010 - link

    Antec used to have a reputation for using low temperature Caps that would fail at actual operating temperatures over time. I had assumed that they stopped doing this, but it is possible that they still use this practice.
  • TGressus - Thursday, July 8, 2010 - link

    In general the more popular ODMs will have a portfolio of designs. Some better than others. In my experience this is where some of the vendors get away with shady practices. Entry level components get hidden within a shiny exterior, marketed as something they are not and no one is the wiser.

    I'd like to see the status quo among reviewers/enthusiasts change to focus on relative comparison, and historical performance of the ODM models inside PSUs. As it stands now we tend to focus on a visual inspection, and essentially a second wave of manufacturer QC testing.

    When we discuss SSD or GPU we consider the chipset first, and the brand second. This enables the community to make better informed decisions, and allows us to steer trends in product development, rather than the vendor hooride that is the PSU market.

Log in

Don't have an account? Sign up now