Advanced Transient Response Tests

In these tests, we monitor the response of the PSU in two different scenarios. First, a transient load (10 A at +12V, 5 A at 5V, 5 A at 3.3V, and 0.5 A at 5VSB) is applied to the PSU for 200 ms while the latter is working at 20% load. In the second scenario, the PSU, while working at 50% load, is hit by the same transient load. In both tests, we measure the voltage drops the transient load causes using our oscilloscope. The voltages should remain within the regulation limits defined by the ATX specification. We must stress here that these tests are crucial since they simulate transient loads a PSU is very likely to handle (e.g., booting a RAID array, an instant 100% load of CPU/VGAs, etc.). We call these tests Advanced Transient Response tests, and they are designed to be very tough to master, especially for a PSU with a capacity below 500 W.





The G-550's capacity may be low, but it performed decently in our Advanced Transient Response tests, especially on the +12V rail where deviations stayed within 1%. However, response time on the same rail was rather low. Voltage drops on all other rails but on 3.3V were also well controlled


Below are the oscilloscope screenshots we took during Advanced Transient Response Testing.

Transient Response at 20% Load

Transient Response at 50% Load

Turn-On Transient Tests

We measure the response of the PSU in simpler scenarios of transient load—during the power-on phase of the PSU—in the next set of tests. In the first test, we turn the PSU off, dial the maximum current the 5VSB can put out, and switch on the PSU. In the second test, we dial the maximum load +12V can handle and start the PSU while the PSU is in standby mode. In the last test, while the PSU is completely switched off (we cut off power or switch the PSU off by flipping its on/off switch), we dial the maximum load the +12V rail can handle before switching the PSU on through the loader and restoring power. The ATX specification states that recorded spikes on all rails should not exceed 10% of their nominal values (e.g., +10% for 12V is 13.2V and 5.5V for 5V).



Performance in the first two tests was very good, but things don't look as rosy in the third and last test because of a short but high section with ripple before the rail's voltage settled down. Thankfully an extreme scenario that may never be reproduced in a real system, we would still like to see better results in the third of the above tests.