A rather large chunk of the PSU’s overall output power falls on the -12V and standby source despite the fact that such a high load capacity of the former rail is not needed in any computer. These 40 watts are a mere 4%, though. Otherwise, I can find no fault with the specifications. The PSU can yield its full 960W across the +12V rail only. The latter is divided into four “virtual” output lines.
There is a small error in the 1200W model’s specs: the +12V rail is specified to deliver 99A or 1188W whereas the maximum allowable load of all the PSU’s main output rails is 1160W. That’s nothing serious, though. It is clear that the PSU can yield its full 1160W across the +12V rail. The label shows you what exactly +12V line powers what, even indicating the color of the appropriate cables (yellow, yellow and black, etc). Such information can be important because the PSU will not work if one individual line is overloaded, even though the whole PSU will still have a large reserve of total output power.
Together with my UPS this power supply worked at loads up to 380W when powered by the mains. When powered by the batteries, this pair was stable at loads up to 340W. Thus, these PSUs have no problems with UPSes.
The +12V voltage is very stable. The +3.3V is less stable but leaves the permissible limits at extreme loads only. The +5V voltage is initially set too low in this PSU and comes close to the bottom limit (4.75V) under high load.
The 1200W model is better in this test as the voltage on its +5V rail is indeed 5V under low loads. However, you can see that the diagram has chaotic color spots instead of sharp transitions between different colors. You will learn the reason for that in the next section.
Output Voltage Ripple
The output voltage ripple of the 1000W model is normal. There are occasional high spikes, but they are not dangerous at all. The oscillogram of the +12V rail is especially good, although this rail has about 90% of the total load.
The high-frequency pulsation of the 1200W unit is almost the same, but do you see that the lines are somewhat slanting?
The reason becomes clear if the oscilloscope’s resolution is increased from 10 to 100 microseconds per division. The outputs based on individual switching DC-DC converters have a strong ripple at a frequency of 3.8kHz which is about as high as the allowable limits. This is why our testbed could not measure the PSU’s output voltages precisely and the cross-load diagram had color spots.
So, I get a feeling that the new DSR series needs more polishing off. At least, both samples I tested have problems you don’t expect to find in such high-class PSUs. The 1000W model has a reduced voltage on the +5V rail whereas the 1200W model has undesired voltage pulsation at the output, which may have a negative effect on the computer’s stability.