Didn't need such a big PSU?

Toggle sidebar Toggle sidebar
I

Iamsoda

Distinguished
Feb 18, 2015
1,712
0
20,310
Well finally got my hands on one of those nifty little power supply gadgets to test watt usage. I plugged it into my pc and started running some test from idle to max benchmarks. Here are the watts used:

Off 2.7-3 watts (pretty sure I have the stupid wifi wakeup magic packet option enabled)
Idle logged in doing nothing: 130 watts
Heavy GPU benchmark: 360 watts
Heavy CPU benchmark: 300 watts

I have my GPU OC to 1.45ghertz
I have my CPU OC to 4.35Ghertz

Granted I did plan on having some overheard for 2 way SLI GPU's if I ever wanted to do so, but really is a 750 watt PSU needed when only 360 max is used? This is power coming straight from the wall measured which means it includes the PSU power wasted. Even with OC I am barely budging the power supply. Was 750 watts overkill?

PS: I do realize I appreciate the fact that I don't hear any coil whine from PSU or any fan noise at all.
 
Solution
InvalidError

Look at GPUs' power consumption analysis in some THG GPU reviews where current draw through each rail gets recorded, you will see that GPUs have a 2:1 or even 3:1 ratio between peak and average power during benchmarks. GPUs and CPUs are not at a steady XX% load during game benchmarks, it varies considerably from frame to frame depending on what is happening between and in those frames.

If you have a 500 microseconds +300W peak on the PSU's outputs and a power meter that does per-cycle averaging, that peak gets reduced to 10W as seen from the input over a 16.7-20ms power line cycle. You...
Sort by date Sort by votes
koffeeshop77

koffeeshop77

Reputable
Jun 9, 2015
1,160
0
5,660
you are looking at it from the wall.
think about lookin at it from inside, the otherside of the box, e.g
open case loop all wires indside the apm msureintool there is prolly a difrent number and inside there is DC not AC so make sure to switch .
and also see how much the GPU uses and then double it seeing you wanted to go sli later on.
 
Upvote 0 Downvote
turkey3_scratch

turkey3_scratch

Expert
Ambassador
Jul 15, 2014
22,630
8
60,960


AC > DC always.
 
Upvote 0 Downvote
InvalidError

InvalidError

Titan
Moderator
May 18, 2007
23,815
5,265
108,890

It depends on the meter and the type of coupling transformer is used. Some designs can do DC, others are AC-only. If the clamp-meter does DCA, it should say something like "Bandwidth: DC-100kHz" somewhere on it or on its spec sheet. Otherwise, it will say 40Hz-100kHz or something else of the sort that excludes DC and very low frequencies.
 
Upvote 0 Downvote
I

Iamsoda

Distinguished
Feb 18, 2015
1,712
0
20,310


That may be true, but I do know one thing there are not more watts being used past the AC outlet inside the computer box. That is I am measuring the max wattage being used by measuring the PSU plug in.
 
Upvote 0 Downvote
P

Plagueis44

Reputable
Jan 21, 2016
6
0
4,510
Just because it's a 750W PSU doesn't mean it will use that. If it draws 360W then you have the 390W head room for adding SLI or other components. Power supply units are actually most efficient running at 50% load. So you have got a good PSU for when you run your system under its larger loads.
 
Upvote 0 Downvote
I

Iamsoda

Distinguished
Feb 18, 2015
1,712
0
20,310


You don't get it do you? I measured the power from the AC cord that is the max the PSU is using.
 
Upvote 0 Downvote
InvalidError

InvalidError

Titan
Moderator
May 18, 2007
23,815
5,265
108,890

Not the max, the average including PSU losses over some period of time which is usually a few AC cycles. The single-cycle or single-alternance maximum will be higher than a longer term average.

The power peaks a PSU needs to cope with can be as short as tens of microseconds and would not register on a Kill-A-Watt or equivalent since it gets averaged and filtered out by the input filter caps and inductors.
 
Upvote 0 Downvote
I

Iamsoda

Distinguished
Feb 18, 2015
1,712
0
20,310


Didn't think about the capacitor surges. Good point, but even so wouldn't a maxed bench mark running at a steady pace hit the PSU up the same?
 
Upvote 0 Downvote
InvalidError

InvalidError

Titan
Moderator
May 18, 2007
23,815
5,265
108,890

Look at GPUs' power consumption analysis in some THG GPU reviews where current draw through each rail gets recorded, you will see that GPUs have a 2:1 or even 3:1 ratio between peak and average power during benchmarks. GPUs and CPUs are not at a steady XX% load during game benchmarks, it varies considerably from frame to frame depending on what is happening between and in those frames.

If you have a 500 microseconds +300W peak on the PSU's outputs and a power meter that does per-cycle averaging, that peak gets reduced to 10W as seen from the input over a 16.7-20ms power line cycle. You still need a PSU capable of delivering that 300W peak over whatever the baseline load is.
 
Upvote 0 Downvote
Solution
You must log in or register to reply here.

TRENDING THREADS

Latest posts

Moderators online

Share this page

COMPANY
RESOURCES
FOLLOW
Top Bottom