Whenever I read a PSU review, I always want to see how efficient it is at the 10-40W DC end as that seems to be where our computers spend most of their time these days.
Would it be possible to just show the results at 20W DC for 115VAC and 230VAC as well as the mid-range test that you do? That would also allow you to see if there are any weird results occuring at low loads when on 230VAC, as well as satisfying my own curiosity as a person in a 230VAC country
Yes, that's doable. OK.
Just as an aside, I predict this year we will start to see the first no compromise gaming computers being assembled that will idle under 25W AC. This should be achievable right now with a well chosen PSU and motherboard with a HD7950 or HD7970 that switches off their fans and use around 1W when entering ZeroCore Power mode. The next evolution of this would be integrating a special core into the GPU that can do basic 3D effects and video decode while consuming 1W max, like the designs being used in the current tablets from Apple and Samsung, etc. Ah well, I can only dream...
I dunno.... maybe <30W, but I can't see <25W.
The most efficient PSU I measured at 22.5W load (Kingwin LZP550 Platinum) was 77.5% -- it drew 29W AC. If we lower the output, efficiency gets worse quickly. For that PSU to draw <25W AC, the DC output would have to be <18W. Is this feasible now with motherboard, RAM, SSD & 1W video card? Maybe. We're also going to start seeing 80 Plus Titanium PSUs some time this year. 90% eff. at 10% load, 92% at 20% load, 96% at 50% load and 91% at full load. That's only for 230VAC; no Titanium defined for 115VAC. But that probably won't help with <20W load -- you need a <200W rated Platinum for any chance of seeing even a 2W AC power reduction at <20W load (compared to that Kingwin). The lowest power retail Platinum is the 500/550W Kingwins, afaik; Dell has a 350W unit but it's an unsuitable noisy 1U server type.