Methodology

To properly evaluate a PSU besides expensive and sophisticated equipment, the person that conducts the evaluation needs to know not only how to operate the equipment, but also have sufficient knowledge about electronics and a PSU's design. The knowledge part is especially crucial since not even the most expensive equipment can deliver a flawless PSU evaluation if the operator doesn't know how to use it properly and what tests to conduct with it. Also, the operator must be able to identify possible problems that can lead to inaccurate results. According to our vast experience so far, a PSU's evaluation in the majority of cases isn't a straightforward process. 

Our evaluations are conducted in higher than 30 °C ambient and up to 32 °C (± 2 °C), to check the PSU's performance under more realistic conditions, since inside a PC chassis the ambient temperature in almost no case will drop lower. Before we start our measurements, we apply an 80 W load to the PSU for 15-20 minutes, to allow it to warm-up properly and reach a steady-state of operation. In addition, our testing equipment also needs a warm-up period to be able to provide the most accurate readings. 

After the PSU's warm-up load period ends, we set the temperature inside our thermal chamber at 30 °C, and we begin the testing procedure, where we set our loaders to auto mode through our Faganas software before trying over a thousand of possible load combinations on the +12V, 5V, and 3.3V rails. This test usually lasts for several hours, depending on the time interval and the load steps that we use. For every evaluation, we keep the number of load combinations higher than 1,450 combinations and through an advanced data interpolation algorithm that we developed, the final result includes 25,000 readings, allowing us to draw the corresponding charts accurately.

Since our breakout board is equipped with a large number of sockets, we can connect up to ten (10) PCIe and two (2) EPS connectors, along with four 4-pin peripheral connectors and a single SATA one. This way we make sure that the voltage drops on cables are minimized even in high capacity units, so our results have increased accuracy. 

The AC source is set to deliver 115 VAC with 60 Hz or 230V AC with 50 Hz frequency. So besides "clean" power, we ensure that in all cases, the registered deviations are kept at very low levels (+-1%). The AC source is powered by a 3 kVA online UPS, which is connected to a 3 kVA transformer. 

For the 5VSB rail's efficiency measurements we set the loading step at 0.05 A and we provide a long enough interval, to allow for the PSU to reach a steady-state in each load level. For the ErP Lot 6 and ErP Lot 3 efficiency requirements, we conduct separate tests since the load step is less than 0.05 A for the two lower load levels. 

For the standby power measurements, we strictly follow the internationally recognized IEC 62301 standard. For measurements with current as low as 1mA external current shunts can be helpful, but their use isn't necessary for IEC 62301 testing. External shunts make the measurement procedure more complex and lead to measurement errors. The best power analyzers, like the ones we use (N4L PPA1530), can test to IEC 62301 by relying only on their internal current shunts. Generally, advanced power analyzers with a large enough dynamic range don't need external current shunts to measure standby (or vampire) power. 

According to the IEC 62301 standard, the standby power measurements should be made with the Total Harmonic Content of the AC source (up to and including the 13^th harmonic) must be less than 2%. In addition, the Voltage Crest Factor should be between 1.34 and 1.49. 

We use our custom program to perform the standby power measurements and to verify our results, we use the Newtons4th's Standby Power application, which is compatible with our power analyzer. You can find more info about the standby power measurement procedure according to the IEC 62301 standard by reading this application note, which is provided by Newtons4th. 

We measure the fan's noise from one meter away, inside a hemi-anechoic chamber whose internals are entirely covered in specialized soundproofing material. Background noise inside the chamber is kept below 6 dB(A) during testing, with humidity being close to 50%. We should stress however that to obtain the fan speed results we apply the same conditions with our efficiency readings (operating temperature >30 °C) and we take the fan's output speed throughout the PSU's entire operating range. This allows us to calculate the unit's overall noise output and provide the corresponding Lambda certification.