# TPU PSU Performance Rating Methodology Explained

Author: crmaris
Date: 2011-11-18 15:18:03

### Voltage Regulation

In an ideal world a PSU would maintain a constant voltage level regardless load, but in real life scenarios there is always a voltage drop on each rail as the load increases. In our tests voltage regulation shows the difference between the initial voltage readings with 60W of load (and not with no load at all since many PSUs don't work well with no load on their rails and some others feature no load protection) and the voltage readings on all rails with full load.

We calculate the voltage regulation rating of all ratings using the following equation:

In the formula above, you can see that 5V and 3.3V have a lower weighting factor and 5VSB the lowest. This is because the 12V line is by far the most important in a system and its stability is crucial. We use the same system in the other formulas below, too.

### Ripple Voltage

Ripple are the AC fluctuations (periodic) and noise (random) found in the DC rails of a PSU. We take the ripple readings of all rails from the 100% load test and combine them in equation shown below.

How well a power supply reacts to sudden changes in load is a very good indication of the unit's power quality. We take the higher deviation each rail registered in the two Advanced Transient Response tests we conduct in order to calculate the performance.

### Turn On Transient Voltage Overshoots

Performance calculation in these tests is rather simple since we use only the rough value of the registered spike, compared to the nominal value of +12V and 5VSB rails. At +12V we use only the higher voltage overshoot and ignore the other one (if there is any of course). To give an example: if the higher registered voltage overshoot at +12V is +12.5V then the spike is 0.5V (12.5V – 12.0V) so we use this value in our equation.

### Efficiency

For must users out there efficiency is one of the most important factors when they look for a new PSU. In our equation we take the mean value of efficiency readings at 20%, 50% and 100% of the unit's maximum rated capacity load, subtract it from 100 (represents the perfect 100% efficiency).

### Capacity

When you directly compare the voltage regulation, efficiency or ripple readings of a 400 W unit to those of a 1500 W model then the comparison is anything but fair.

For example, when looking at voltage regulation at 60 W load, the regulation range for a 400 W PSU is 340 W (= 400 W - 60 W). However, when testing a 1500 W model, it is much bigger: 1440 W (= 1500 W - 60 W). With such a huge difference it is natural that the smaller capacity unit will most likely register much better voltage regulation, or ripple. So we added a normalization that would set things right. This factor is directly derived from the rated power of each unit.

### Final Performance Rating Function

Now that we have calculated all six above factors it's time to combine them to calculate the final performance rating. We take 100 as the perfect score and start subtracting results (1)-(5) and finally take into account the capacity (6)

The result of is the unit's final Performance Rating.

### Performance Rating Calculation Example

For this example we will use Seasonic's flagship, the P-1000 PSU:

(1) 1.41 + 0.5 x 1.43 + 0.5 x 2.24 + 0.25 x 3.33 = 4.0775
(2) 0.1 x (24 + 0.5 x 12.5 + 0.5 x 14.9 + 0.25 x 13.6) = 4.11
(3) 0.2 x (1.34 + 0.5 x 1.82 + 0.5 x 2.57 +0.25 x 2.37) = 0.8255
(4) (0+0.5) = 0.5
(5) 0.5 x (100 - 1/3 * 90.81 + 1/3 * 92.79 + 1/3 * 91.16) = 4.206667
(6) 1000 / 50 = 5

Performance Rating = 100 – 4.0775 + 4.11 + 0.8255 + 0.5 + 4.206667 + 5 = 100 – 13.569 + 5 = 91.280333

### Conclusion

We hope that by now you understood the methodology we follow to calculate the performance rating of each PSU. We tried to make it as clear as possible and we are sure that the math is fairly easy to follow. Your comments are welcome!

Page: