High Wattage PSUs - Do they consume more power?

[Andi]

i liked the idea of the video, but the execution was awful. Using a 450W bronze psu was just not a good idea. You can buy a 400W platin rated psu, or at least a gold. You should talk to Corsair, Seasonic or Cooler Master, and do that kind of video with one specific lineup of psus in mind.

[Kuvinen]

Maybe luke can better do a video on  efficiency from a psu..

Like how do you know what the sweetspot is for a psu for your system..

 

Atm when i think about getting a new psu i just first look at is it 80plus gold or higher.?? 

A 80plus bronze psu i would never buy... 

 

I really think you can better check it out with the same efficiency label and also you can check out with how much load it would be better or worse?

[Markmjb]

Decent video, the summary at the end was too vague, if I were to summarize it :

Different power supplies take different wattage from the wall in an equal scenario, but since most power supplies nowadays are of a high efficiency rating, the effect is negligible.

You may have found the right answer to the wrong question though... you could take this topic way further. 

 

I took some further time to calculate some stuff: 

The average kwH price here is €0,22 in 2015.

Kwh year assumes 8 hrs a day = 2920 hours.

€ every 100 hrs is the cost of that thing every 100 hrs.

Let's say I am making your testbench on my wish list, and want to get the best investment of my money. The point I am calculating is where the energy cost and the cost of the power supply are at the same level. 

 

If I have the 450W:

750W I need to make €51. This takes 4 years.

1200W: I need to make €263.This takes 31 years

1500W: I need to make €361. This takes 23 years. 

 

If I have the 750W:

1200W: I need to make €212. This takes 55 years.

1500W: I need to make €296. This takes 91 years.

 

If I have the 1200W:
1500W: I need to make €82. This takes 12 years.

 

It's fun and insightful to calculate this stuff. The 8 hrs is a low estimate, and if I have that computer on twice that much in a day, it would only take 2 years to justify the upgrade to a 750W. So that choice would make sense, it would leave me with upgrade room, and even if I don't choose to add an additional gpu, I'd still save money on the long run. 

 

However much hours a day, if you think you are gonna run heavy sli/cf, you might consider the 1200W. The 1500W makes no sense from a financial perspective whatsoever.

 

That's kind of the point I missed from your video. If you were to shop for a power supply just for the sake of efficiency, you should really take these numbers in account. The higher lvl wattage psu's sound great on paper, until you calculate how long it would take to justify its purchase. 

People really should make this sort of calculations with psu's of similar wattages. Maybe that's a good topic for another video.Take 10 650/750W power supplies, and see what their value is calculating above things.

 

[-Old-Tom-]

9 hours ago, linuxfan66 said:

according to this and other research combined in my head.  240v has less resistance to deal with due to half the current per voltage. Second 240v conversion is apparently designed more efficient by requirement. But i could be wrong

It also has to do with how bridge rectification and diodes work. diodes will always clip the same amount of voltage off the input. ie taking 1.2 V off 110 V = 1.1% power loss whereas 1.2 V off 230/240 V = about 0.5% power loss. as its a bridge rectifier you have to double that which makes 110 V losses around 2.2% compared to 230s 1% ish.

so you can expect power supplies running off 110 volts to be about 1% less efficient than if they were running the same system using a 230 V supply and that's only looking at one small part of the conversion process. 

[Marinatall_Ironside]

4 hours ago, Enderman said:

He did, but that completely made all the tests irrelevant since they dont answer the question of "high wattage PSUs do they consume more power"

Because if you compare a 600W PSU with a 1200W PSU with the same efficiency, at idle the larger PSU will consume more power

Also if you compare those two PSUs with a PC that only uses 300W at load, then the larger PSU will still use more power

 

SO basically the video was "here are these tests which dont show any pattern at all" and "bigger PSUs may or may not consume more power" and "go learn about efficiency curves yourself"

 

And I'm pretty sure that a lot of the people on youtube are not going to bother to research how efficiency curves and % load affects the power consumption of a PSU

If you've listened the end of the video, Luke explicitly said that you should find the sweet spot of the efficiency curve and go with it, which validates what I have highlighted in teal. Yes a bigger power supply will draw more power at load because it's not at the 50% mark, which is where the efficiency curve peaks. So if you have a system that draws 300 watts at full load, you should be getting a 600 watt power supply (with 80 Plus Gold preferably) so that way, at full load, you're at 50% of the rated power capacity of the PSU.

[Enderman]

7 minutes ago, JurunceNK said:

If you've listened the end of the video, Luke explicitly said that you should find the sweet spot of the efficiency curve and go with it, which validates what I have highlighted in teal. Yes a bigger power supply will draw more power at load because it's not at the 50% mark, which is where the efficiency curve peaks. So if you have a system that draws 300 watts at full load, you should be getting a 600 watt power supply (with 80 Plus Gold preferably) so that way, at full load, you're at 50% of the rated power capacity of the PSU.

I dont think you understand how this works...

 

Efficiency curve is only half the story here

Efficiency rating moves the entire curve up or down

And since the tests done had both variables, that doesnt answer the question in the title

Foe example, not all 1200W PSUs will be 80+ platinum and not all 600W PSUs will be 80+ bronze

 

Maybe you've never done any kind of scientific testing or research, but there needs to be only ONE variable to get any meaningful/useful data out of the result

[Marinatall_Ironside]

2 hours ago, Enderman said:

I dont think you understand how this works...

 

Efficiency curve is only half the story here

Efficiency rating moves the entire curve up or down

And since the tests done had both variables, that doesnt answer the question in the title

Foe example, not all 1200W PSUs will be 80+ platinum and not all 600W PSUs will be 80+ bronze

 

Maybe you've never done any kind of scientific testing or research, but there needs to be only ONE variable to get any meaningful/useful data out of the result

I understand how it works. Do you think I have the disposable income to actually conduct any experiments on power supply efficiency and how power is drawn? Hell no! And of fucking course I know not every 1200 watt or 600 watt PSU will be 80 Plus Platinum or Bronze, but when you move up to those levels, they tend to hit that 80 Plus Platinum efficiency dead-on because manufacturers doesn't skimp on build quality, component quality, component choice, and circuitry design, whether they're modular or not, but manufacturers almost always goes for the fully modular design. Also with Bronze in mind, the majority of PSUs on the market are 80 Plus Bronze, so finding one isn't all too difficult, even for lower-end units (excluding the God-forbidden Corsair CX PSUs).

 

You also know it's never good to run a power supply at almost full load, right? I'd much rather go overkill and use up half of 1.2 KW of a power supply than to cut it close with a 650-700 watt PSU because at 1.2 KW with 80 Plus Platinum or Gold, you'll be at 50% of the load, which is where the efficiency peaks at. At low and max loads, you'll see that you don't get that much efficiency at all. And when you do run a power supply at near-full load, the PSU will not last as long due to heat output when stressed to those levels.

[dragoon20005]

11 hours ago, aw_ said:

C19 is not proprietary https://en.wikipedia.org/wiki/IEC_60320#C19.2FC20_coupler

correct

11 hours ago, WillG said:

I don't think Luke has worked in a data center or with blade chassis. 

erm they have blade servers storing their videos and they bound to have UPS powering them

 

it because of the 120V needs a lot more current around 12.5A to get 1500 watts, normal computer power cable which are rated at less than 8A will melt with the huge current passing thur the wires.

 

 

 

well the point of this video is to point out the stupid myth that buying a bigger wattage PSU = higher power bills

 

which is f***ing bulls***

 

who ever tell me that gets huge slap from me

 

 

back to the video

 

Luke did touch a bit on efficiency of the PSU from Bronze, then jump to gold, followed by Platinum and finally Titanium

 

there was a small number of Silver efficiency PSU but i guess they were phased out for Gold standards minimum for higher wattage PSU

 

I think not many end user will understand how to reach the peak efficiency curve just by looking at the graph on the box or the website. This is really true for first time PC builders.

 

also the main issue with power efficiency works best with 230 voltage

 

Canada uses 120 volts similar to US so the current passing thur the PSU is way higher which leads to power wastage in a form of heat.

 

it good your are testing the same hardware and games to get the power draw.

 

I am sure you could do the test with more similar PSU commonly found in PC builds like the CX series

 

yes GUYS put down your pitch forks

 

let Luke do an honest test on how well these PSU are before we pass judgement that they should not be touch.

 

take it from CX430/M then followed by the CX500 and CX600 and then the CX750

[Ranggon]

Luke used a 450w PSU for a gtx980ti, in theory my corsair vs450 PSU should be able to power my gtx 970 safely right?

[lo_pan_2000]

I'm looking to upgrade my power supply to something that will work well into the future. Can someone tell me if a higher rated, more efficient power supply is more reliable? Will it last longer before it dies than a lower rated PSU? I guess my question is, does the higher rated power supply have to work harder to be more efficient and will that put more wear and tear on it and give it a higher chance of failing faster than a lower rated supply?

[Markmjb]

2 hours ago, lo_pan_2000 said:

I'm looking to upgrade my power supply to something that will work well into the future. Can someone tell me if a higher rated, more efficient power supply is more reliable? Will it last longer before it dies than a lower rated PSU? I guess my question is, does the higher rated power supply have to work harder to be more efficient and will that put more wear and tear on it and give it a higher chance of failing faster than a lower rated supply?

Well, remember that a power supply is  connected to hardware that is way more expensive than the thing itself, like motherboards, graphic cards and so on. So you should choose a respectable brand for your power supply. The rating itself doesn't matter that much, since most of the good brands already have like 5 or 6 years warranty. The quality of those PSU's is good enough that they can have that kind of warranty. 

[NumLock21]

On 2/15/2016 at 9:19 AM, aw_ said:

C19 is not proprietary https://en.wikipedia.org/wiki/IEC_60320#C19.2FC20_coupler

Yep, they're not proprietary. Even some Dell XPS gaming system uses the C19 power cord.

22 hours ago, Andi said:

i liked the idea of the video, but the execution was awful. Using a 450W bronze psu was just not a good idea. You can buy a 400W platin rated psu, or at least a gold. You should talk to Corsair, Seasonic or Cooler Master, and do that kind of video with one specific lineup of psus in mind.

IMO Luke did a better real world test by using different rating PSUs, because who in the world is going to spend $100 on a 450w Platinum rated PSU? I doubt anyone here would spend that much on it, unless they're obsessed with their power bill, where a rise in a penny will makes them go crazy.

FSP 450w Titanium cost $99.99

EVGA 750w Gold cost $99.99

Between the 2, I'll grab the EVGA one for more headroom. Never really get the point of buying a psu, that's borderline with the components that they have. Like whole system needs 500w and they buy a 500w psu, instead of buying a 600 or 650w.

 

8 hours ago, lo_pan_2000 said:

I'm looking to upgrade my power supply to something that will work well into the future. Can someone tell me if a higher rated, more efficient power supply is more reliable? Will it last longer before it dies than a lower rated PSU? I guess my question is, does the higher rated power supply have to work harder to be more efficient and will that put more wear and tear on it and give it a higher chance of failing faster than a lower rated supply?

I'm currently still running on my Antec Quatro 850w PSU, that's only 80 Plus rated. It was one of the 80 Plus rating psu of that time. "Olympic" style rating psu didn't exist yet, until years later. The PSU is in my system for 9 years and it's still working great.

A newer version is Bronze rated with 82% efficiency, the original is just 80% efficient. When buying a PSU, set a budget on how much your willing to spend, and then look at the tier ranking.

 

Review

http://www.jonnyguru.com/modules.php?name=NDReviews&op=Story&reid=58

[-Old-Tom-]

10 hours ago, lo_pan_2000 said:

I'm looking to upgrade my power supply to something that will work well into the future. Can someone tell me if a higher rated, more efficient power supply is more reliable? Will it last longer before it dies than a lower rated PSU? I guess my question is, does the higher rated power supply have to work harder to be more efficient and will that put more wear and tear on it and give it a higher chance of failing faster than a lower rated supply?

If your looking at making your power supply and the rest of your system last longer i would suggest getting a line interactive UPS. This not only protects your system from sudden power failures, most of them have good surge ratings, and the line interactive UPSs mean that your system is always getting a nice consistent sine wave in as opposed to the raw wall power which can have funny wave forms at times and the voltage moves around a fair bit depending on what your neighbors and the rest of the power grid is doing.

[Andi]

On 16.2.2016 at 6:37 PM, NumLock21 said:

IMO Luke did a better real world test by using different rating PSUs, because who in the world is going to spend $100 on a 450w Platinum rated PSU? I doubt anyone here would spend that much on it, unless they're obsessed with their power bill, where a rise in a penny will makes them go crazy.

FSP 450w Titanium cost $99.99

EVGA 750w Gold cost $99.99

Between the 2, I'll grab the EVGA one for more headroom. Never really get the point of buying a psu, that's borderline with the components that they have. Like whole system needs 500w and they buy a 500w psu, instead of buying a 600 or 650w.

Real world? If you run a normal overclocked quad core with a 980Ti you will bearly suck 400W from the wall. If you don't plan to go SLI/CF anything over a 550W psu is just a waste of money and electricity. And please don't tell me, you might want to go SLI in the future. Noone thinking that ever did it.

 

Here in Germany you have to pay nearly 0,3€/KWh, playing 2h every day with a normal PC, using maybe 300W, you will save 5-7€ a year. The difference between a good bronze and a good gold psu is around 30€. In your use case of 8 years you could have saved 25€ already...assuming you PC only runs 2 hours a day, but it could be much more if you use it more.

 

[mooseman]

@SlickDid you happen to record the idle power draws? As others have mentioned, idle is actually more important when it comes to figuring out power usage costs.

[papapoi]

what is the tester called that luke used to test the power draw, im looking to get one, my son has a 725 sentey PSU which has no 80+ rating but had good reviews for $45, im wondering how much power is that psu wasting if any, considering he only has a 660ti with and i5-6400, im already paying over $300 USD a month for electric(the price i have to pay for living in this so-called paradise called Hawaii)

[obsidian1200]

1 hour ago, papapoi said:

what is the tester called that luke used to test the power draw, im looking to get one, my son has a 725 sentey PSU which has no 80+ rating but had good reviews for $45, im wondering how much power is that psu wasting if any, considering he only has a 660ti with and i5-6400, im already paying over $300 USD a month for electric(the price i have to pay for living in this so-called paradise called Hawaii)

Not sure exactly which one Luke is using, but you can pick up testers that do the same thing from most hardware stores (though idk about stores in Hawaii). One of the more popular ones is called the Kill-A-Watt. 

 

Edit: Link to kill-a-watt on amazon: http://www.amazon.com/P3-P4400-Electricity-Usage-Monitor/dp/B00009MDBU/ref=sr_1_1?ie=UTF8&qid=1456297304&sr=8-1&keywords=kill-a-watt

[papapoi]

1 minute ago, obsidian1200 said:

Not sure exactly which one Luke is using, but you can pick up testers that do the same thing from most hardware stores (though idk about stores in Hawaii). One of the more popular ones is called the Kill-A-Watt. 

thanks found it on amazon for $20, im sure its $50 here in Hawaii