PSU Rating vs. Energy Consumption

_SBradley_

_SBradley_

_SBradley_

Associate
Joined
4 Jul 2009
Posts
736
Location
Perth, WA, Australia
Hi guys!

I think I know the answer to this, but I would like to check.

Would a lower-rated PSU and a higher-rated PSU (from the same manufacturer, with the same internal architecture) consume the same amount of energy, given the same system to run?

In other words, would you pay extra on your energy bills, if you were to "overdo" it on the PSU? Or is it just that the PSU would have power in reserve, were the system ever to require it?

Thanks! :)
 
No, they will both draw the same from the wall soclect assuming they are just as efficient. Though PSU are meant to be most efficient when running at around 80% of capacity.
 
Yeah smaller PSUs will generally use slightly less power as they'll be closer to their most efficient point. If you're only using 200W on a 1KW PSU the efficiency will be a fair bit worse than if you were using a 400W PSU and even then the 400W PSU will be pretty far from its ideal.
 
I wonder, if the PSU is not near its most efficient point, will the PSU also be under more stress? That is, is it likely to go "pop" sooner if it's being underutilized?
 
If you look at the some of the efficiency vs loading graphs on the 80plus site, it seems that peak efficiency is typically at around 50% loading, but the graphs are really fairly flat between 25% to 100% loading. Below 25% loading, the efficiency starts to drop off significantly.

So speccing a 1000W PSU for the average PC or server that normally draws under, say, 200W is completely pointless, and will cost more in electricity as well as the power supply itself.

I can't see why running a PSU at a low loading would have a significant effect on reliability (at low load it would be dissipating less heat and run cooler than at full load).
 
READ THIS: CLICK HERE
jonnyGURU said:
On the consumer level, there really isn't any options to the Kill-A-Watt and other devices like it. And I can't see somone paying hundreds of dollars for equipment just to measure power consumption unless the review power supplies or do something for a living that would require that.

Some people say adjust for efficiency (80%) and then consider keeping the load within 50 to 75% of whatever PSU's capability. So for example, 300W from the wall, x 80%, 320 WDC... 500W to 550W PSU. But to take into consideration inaccuracy of the Kill-A-Watt and immeasurable transients, I would just suggest double whatever the Kill-A-Watt is reading as a PSU recommendation.

So 300W from the wall, 600W PSU. Just to be "safe". You're still not suggesting 1600W PSU's for people with 8800 GT cards (one side of the spectrum) but your not telling people to buy 380W PSU's that will die in a year's time due to stress (other side of the spectrum.)
Click to expand...
You get a lot of people on here reccomending using power supplys 400W etc & their like using 350W at load etc, just because a PSU is within its wattage doesnt mean it should be reccomended.

The guy who wrote that above is a PSU expert, worked for various PSU companys, currently works at BFG Technologies as a product manager so he knows what he's talking about.

A PSU is most efficient usually at 50% load.
 
Last edited:
Nelly said:
A PSU is most efficient usually at 50% load.
Click to expand...
Only some later ones.
(whose low load efficiency is tried to keep decent)

And point where efficiency differences really mean something is idle consumption of PC where PCs spend most of their time.
And that's sizing method will surely drop that point under 20% load most of the time.
 
Superewza said:
It's not 50, it's around 80%. But it differs a little for each PSU, obviously.
Click to expand...
4 totally random PSU tests from the past 2 years of reviews. Theirs not much in it mostly but it definitly peaks @ 50%.

Even 80plus website will show the same, check their graphs.

psu1x.jpg
psu2x.jpg


psu3.jpg
psu4.jpg
 
In terms of ball park figures, which are realy the only figures you can use due to the amount of variables involved from system to system, I would say nelly is pretty much bang on, over 80% load is generally where efficiency starts to drop off, NOT where it is most efficient....

Basicly you should aim to have at least 20% (at load) overhead. On the other side of the coin, running a PSU at an average 20% of its capacity is probably going to be just as inefficient.

Not to be confused with the 80+ certification, which is to do with power wasted as heat during the conversion from 240vAC to 12vDC, this just means that on average, the PSU will waste 20% or less power (as heat) during the conversion.
 
You must log in or register to reply here.
Back
Top Bottom