For PSU Test Rig V.3, we revised our testing equipment and methodology to improve the accuracy of the load testing and efficiency calculations for our power supply reviews. As part of this revision, we re-tested most of the power supplies on our Recommended PSU List using the new methodology. The new results are reported in this article along with the original results.
October 20, 2005 by Devon
Cooke and Mike Chin
Recently, we revised our
testing equipment and methodology in PSU Test Platform V.3 to improve the accuracy of the load testing and efficiency calculations
for our power supply reviews. As part of this revision, we re-tested most of the power supplies on our Recommended PSU List using the new methodology. The new results are shown below along with the original results. This
data will also be added to the relevant reviews.
In general, the biggest errors occurred at higher
output levels, typically above 200W output. At lower output levels, the efficiency error
was typically no more than one or two percentage points.
The relative efficiency of the tested power supplies has not changed.
If the tested PSUs are ranked by efficiency, the rankings remain the same whether we use the original results or the new results. You can see all of the gory details below.
Models are listed alphabetically. The model name links back to the original review.
| CORRECTED PSU OUTPUT & EFFICIENCY | ||||||||||||
| Model | Target Output | 40W | 65W | 90W | 150W | 200W | 250W | 300W | 400W | 500W | 600W | |
| Antec Phantom 500 | Measured Output | – | 67.2W | 92.8W | 152.2W | 198.8W | 251.9W | 299.1W | 398.9W | 499.9W | – | |
| Efficiency | Corrected | – | 74.7% | 77.3% | 81.0% | 82.1% | 82.6% | 81.9% | 81.4% | 78.1% | – | |
| Original | – | 78% | 81% | 85% | 88% | 90% | 89% | 88% | 88% | – | ||
| Measured Output | 42.1W | 63.6W | 86.7W | 149.3W | 198.5W | 254.5W | 299.4W | 447.7W† | – | – | ||
| Efficiency | Corrected | 64.8% | 73.1% | 74.7% | 78.1% | 79.7% | 79.1% | 78.4% | 75.1% | – | – | |
| Original | 65% | 71% | 75% | 78% | 80% | 78% | 77% | 77% | – | – | ||
| CoolerMaster Real Power 550 | Measured Output | – | 67.6W | 91.6W | 152.0W | 201.6W | – | 300.3W | 400.1W | 551.7W‡ | – | |
| Efficiency | Corrected | – | 76.9% | 79.0% | 81.3% | 81.3% | – | 80.1% | 77.5% | 74.7% | – | |
| Original | – | 73% | 75% | 77% | 81% | – | 82% | 82% | 82% | – | ||
| Enermax Noisetaker EG701AX | Measured Output | – | 66.3W | 94.2W | 151.5W | 198.2W | – | 297.7W | 448.0W† | – | 599.1W | |
| Efficiency | Corrected | – | 70.5% | 74.1% | 77.7% | 78.3% | – | 78.5% | 76.3% | – | 73.3% | |
| Original | – | 71% | 74% | 77% | 80% | – | 82% | 81% | – | 79% | ||
| FSP Zen FSP300-60GNF | Measured Output | 42.1W | 62.8W | 92.1W | 148.0W | 194.5W | 252.2W | 296.0W | – | – | – | |
| Efficiency | Corrected | 76.6% | 80.4% | 83.0% | 84.6% | 84.6% | 83.6% | 82.9% | – | – | – | |
| Original | 77% | 80% | 82% | 84% | 87% | 87% | 88% | – | – | – | ||
| Seasonic 400HT Active PFC F3 "80 Plus" | Measured Output | 42.1W | 63.6W | 90.2W | 153.5W | 197.2W | 251.2W | 297.9W | 401.3W | – | – | |
| Efficiency | Corrected | 76.6% | 81.5% | 82.8% | 85.3% | 85.3% | 84.3% | 83.9% | 82.6% | – | – | |
| Original | 74% | 78% | 81% | 83% | 87% | 88% | 89% | 90% | – | – | ||
| Seasonic S12-500 / 600 | Measured Output | – | 65.3W | 89.7W | 148.7W | 198.5W | 249.5W | 300.2W | 399.1W | 500.7W | 602.3 | |
| Efficiency | Corrected | – | 75.1% | 78.0% | 81.2% | 82.0% | 81.8% | 81.1% | 79.0% | 79.7% | 78.9% | |
| Original | – | 77% | 83% | 85% | 87% | 87% | 87% | 87% | 86% | 83% | ||
| Seasonic S12-430 | Measured Output | – | 65.8W | 90.9W | 151.7W | 203.6W | 261.7W | 299.3W | 427.2W* | – | – | |
| Efficiency | Corrected | – | 78.3% | 80.5% | 81.6% | 81.8% | 80.5% | 79.6% | 75.5% | – | – | |
| Original | – | 77% | 78% | 79% | 82% | 82% | 80% | 80% | – | – | ||
| Silverstone ST30NF | Measured Output | – | 65.0W | 93.6W | 149.6W | 200.5W | 251.5W | – | – | – | – | |
| Efficiency | Corrected | – | 75.6% | 78.7% | 80.8% | 80.5% | 79.6% | – | – | – | – | |
| Original | – | 76% | 77% | 78% | 81% | 80% | – | – | – | – | ||
| * Target output was 430W | ||||||||||||
Several general statements can be made about all of this:
1. Highest accurate measured efficiency is 85.3%. At this point in time — October 20, 2005 — the highest efficiency measured with our new, more accurate PSU test platform is 85.3%, reached at 150W and 200W output loads with the Seasonic 400HT "80 Plus" approved variant.
2. All previous SPCR-reported PSU efficiencies greater than 85% are simply wrong. We apologize for the unintentional misinformation. We and our equipment both made errors. Our new testing system is accurate within ±1%, as far as we can determine, but the results are always subject to sample variances, which are impossible to eliminate.
3. Only two tested models approached 85% efficiency: The aforementioned Seasonic 400HT "80 Plus" and the fanless Fortron Zen, with peak efficiencies of 85.3% and 84.6%, respectively, reached at the same 150~200W output.
4. A handful of models reached >80% efficiency. Ranking them by efficiency is difficult and depends on which power output level is used. So in simple alphabetical order, here are the other recommended quiet models that achieved >80% efficiency at one output level or more.
- Antec NeoHE 430 (second sample reached 80.3% at 150W load)
- Antec Phantom 350 & 500
- CoolerMaster Real Power 550
- Seasonic S12-430, S12-500 & S12-600
- Silverstone ST30NF (fanless)
5. Small efficiency differences of 2% or less mean very little for most applications and users. The Energy Star division of the U.S. Environmental Protection Agency (EPA) has complied extensive statistics on computer usage. Some of this research is cited on page 7 of the downloadable report, Assumptions for Energy Savings Estimates. (For more information about Energy Star’s proposed changes to PC energy guidelines, please see A New Energy Star… in 2007.) Using statistics from this report, it can be calculated that:
- Average computers in the workplace are in idle mode about 95% of the time that they are on, and at maximum load about 2.5% of the time. (The remainder is in sleep mode).
- Average home computers are in idle mode about 83% of the time that they are on, and at maximum load about 14% of the time. (The remainder is in sleep mode).
Consider the above information with the results reported in our article, Power Distribution within Six PCs.
- Idle DC power delivery for six varying systems ranged from a low of 31W to a high of 83W. A 2% difference in PSU efficiency, say between 78% and 80%, would mean about 1W difference in AC input (or heat) for 31W load, and under 3W difference for 83W load. These difference are trivial and irrelevant for heat dissipation or noise.
- Maximum DC power ranged from 82W to 183W. A 2% difference in PSU efficiency, again, between 78% and 80%, would mean <3W difference in AC input (or heat) for 82W load, and just over 6W difference for 183W load. The 6W difference at full load means the PSU is generating this much more heat, but it’s small enough that other differences in cooling (such as the size of the PSU heatsinks, its internal airflow, ambient room temperature and overall system airflow) could easily swamp the theoretical advantage of a 2% more efficient PSU.
Finally, while higher power efficiency by itself is considered noble these days, remember that the primary reason for SPCR’s long-standing interest in this matter is that higher efficiency can allow a PSU to be cooled effectively with less airflow, which means less noise. But high efficiency in a PSU does not always equate low noise. A noisy PSU with a bad fan and poor controller but high efficiency still does not make our recommended PSU list.
* * *
