PSU Efficiency Makes a HUGE Difference

[MtnHermit]

Reading some of the posts of this forum I had an inkling that efficiency was important, but prior to my test today I always thought a small PSU operating in its power range trumps efficiency, not even close.

Code: Select all

      PSU          FF       No Load         Idle     Comments
Seasonic 180W     SFX        7W              44W     SS-180SFD   
Sparkle 200W      SFX        9W              45W     FSP200-50SNV
Antec EW 380W     ATX        6W              34W*    EA-380 80+
TT 420W           ATX       14W              45W     Thermaltake HPC-420-102DF
Bestec 250W       ATX        6W              39W     ATX-250-12Z Compaq OEM

All the PSU's were connected to a i3-530 on a Gigabyte microATX MB w/4GB DDR3 and a notebook HDD.

The Seasonic is a $14 OEM pull from eBay, the Sparkle came with my Athenatech MicroATX Case, and the Antec is an Earth Watts 80+ 380W PS. No load is everything disconnected with the green wire shorted to ground (black), a PSU is never used this way, only a test.

This has been a real eye opener for me and now I know where to go next since the Antex ATX PSU won't come close to fitting into the Athenatech case. I could easily use it external, the cables are more than long enough, but kludge.

Edit: * The ODD was disconnected, now retested. Added two more PSU

[ntavlas]

It`s true that better overall efficiency can make up for not working in the optimal power range.

This is mostly because of the different tech used in each psu though. A modern 80plus bronze psu like the seasonic 1u @250 watts would beat any 400+ watt psu at those power levels.

Another member has tested a dell 80+ gold low wattage model that even managed to beat the pico psu.

[MtnHermit]

A modern 80plus bronze psu like the seasonic 1u @250 watts would beat any 400+ watt psu at those power levels.

I've not seen that model, is it only available in Greece?

The Dell RM112 is certainly impressive, however its unusual shape makes it a poor fit for my case. A SFX/microATX is what I need.

[ntavlas]

I've not seen that model, is it only available in Greece?

I wish, I got it from the UK. Mine is actually an older version 80+ "only".
While I was looking for more info on it I noticed that seasonic have a new, 80+ bronze revision.

http://www.seasonic.com/product/ipc_1u.jsp

[speedboxx]

Wow Im surprised the EA380 draws so little power at low loads. From the 80plus charts, I would estimate the EA380 to be less than 75% efficient at the low loads you tested at. So the other power supplies must be even worse at efficiency...perhaps less than 60%. For such a low load, I wonder if there would be significant improvement in something like the gold rated Dell RM112 vs the Earthwatts 380.

[MtnHermit]

I wonder if there would be significant improvement in something like the gold rated Dell RM112 vs the Earthwatts 380.

I would certainly expect an improvement, but significant???

[audiojar]

Some PSUs will have a really wide range of peak efficiency, some will drop off fast outside their optimum range. You can't really say whether a 350W will be less efficient at low power than a 150W without looking at both efficiency curves.

Also, the way we talk about efficiency, saying that something is 80% efficient for example, can be misleading. If you hypothetically had a 98% efficient PSU boosting it to 99% would actually be a doubling of it's efficiency, even though you are raising the rated efficiency by 1%. The higher you get the more you actually benefit from 1% more, so actual power savings can be a lot even if the difference in rated efficiency seems small.

[ntavlas]

@audiojar

When going from 98 to 99 percent you are not doubling the efficiency but you are indeed halving the wasted energy, I think that is what you meant to say?

When you look at the total power needed though the savings are rather small as a percentage even though you`ve halved the wasted power in the psu.