Roundup: 12 Gaming Power Supplies Compared

[nirutbs]

Stupid recommmmm..go and read at www.johnnyguru.com

 

[Scott2010au]

[citation][nom]Olle P[/nom]I fail to see how a low airflow temperature is good. To me it shows that either the fan is running too fast, producing unnecessary noise, or the really hot spots don't get sufficient cooling (with most of the air passing through the PSU without touching any heated surface).[/citation]

You are very wise!

 

[Proximon]

I thought this article improved on the last one, and for that I'm grateful.

Your noise, ripple, and efficiency results seem to be right around what others have found for those units they have tested. And yet, your conclusions vary a bit. This is mostly due to incomplete tests that fail to take in all the important criteria.

So let's get into testing methodology, again.
- You need real world stress testing. This is vital. Voltage readings and ripple must be done at conditions that match what could happen to a PSU in the field. That means gaming with a 30C ambient temp while drawing 40C air from an under-ventilated case interior. 40C is the minimum, not the (likely) 24C of your open-air test bench. This alone will dramatically change your results. (Did you know that some better PSUs actually improve in efficiency when warm?)
- Where are the voltage regulation figures? You must have these and yet you fail to mention them in your article and fail to include them as criteria. Even novice, completely untrained reviewers know that voltage regulation is important. Stating that voltage regulation is within ATX spec is not in any way good enough, as the spec is somewhat loose.
Those are the two vital points. Perhaps you are time limited but those are vital.
There are also other important steps that do not depend on lab time. Interior shots (at least one) and a mention of the platform used. The actual OEM stated.

Here are easy things to list that do NOT require a lot of training and no de-soldering:
Fan type and manufacturer.
Primary capacitor(s) rating and brand.
Any impressions of workmanship. Soldering, etc.
Cable type and length.

 

[Scott2010au]

[citation][nom]Gulli[/nom]I'd rather see a comparison between PSUs at the lower end of the spectrum. I can't imagine any of these 750-850W monsters breaking a sweat even with dual high end GPUs. It's uch more interesting to know what's the smallest PSU that can do the job for most desktops. For example: I've been running an HD 5870 with an overclocked Core i7 920 for over a year now on a Corsair 520W PSU, with no problems whatsoever. But if I were to ask on the forums what PSU I would need for my setup most people would recommend 650W, at minimum. So please do a comparison that shows how few watts is enough for single GPU setups and which PSUs are reliable enough.[/citation]

Amen to that!, Name them and shame them!

 

[Scott2010au]

OMG, WTF?, BBQ!

A 'typical desktop PC' as taken from the images uses only 50 watts.

Yeah, for a laptop maybe, but not a typical desktop PC.

Your numbers for each scenario are completely off!

Heck a low power PC uses about 48 watts.

A *typical* PC, with integrated video and one HDD uses between 120 and 160 watts.

Should I keep reading this article, or not? (Was it paid for by Corsair? for instance)

 

[Scott2010au]

The images are not sized correctly for the web page, although they can be clicked on for more information, then right clicked and viewed at a 1:1 ratio (which is readable, thank God).

The article isn't as bad as some of the comments make out, it just lacks a certain level of professionalism overall.

The desire to investigate is there, just the focus is on the wrong details (just a few efficiency graphs).

Something major that is +50% / -33% off is called minor, and something minor that is < +5% / < -4% off is called a major weakness.

That, I suspect, is why you've received so much negative criticism Patrick.

It is much better than previous PSU reviews at Toms Hardware, if the trend continues then I think you may become a crowd pleaser.

 

[Scott2010au]

Is there a table which focuses on the ripple and noise measurements?

All the statistics from the bottom of the images, just in tabular text table format for each unit.

 

[blackmancer]

if u look at the corsair, it has better efficiency than the coolermaster yet you recommend the coolermaster only because it didn't break it's efficiency promise on the packaging. i admit it is disappointing when a manufacturer claims something they dont deliver, but it is still a better PSU. you're comparing apples to oranges in a because they aren't ALL 80plus GOLD or SILVER or whatever u want to choose. but i'd take the corsair anyday - 7yrs warranty, seriously, when do u ever get that kind of backing for a product?

 

[mariushm]

So, since I cannot read. Why don't you tell us all what test the Seasonic didn't pass? I quote directly from the article, "The bottom line is fairly positive with the Seasonic X-750 scoring great results in our efficiency tests. The supply is not without weaknesses, though. Seasonic is generous with its eight Molex and nine SATA connectors, as well as with the three CPU connectors." No mention of what those weaknesses were. Anyone who agrees with the findings in this article either do not know better or blindly support anything Tom's Hardware writes. I've written a final complaint/plea to Tom's in the Site Feedback/Website Opinions section. I am hoping to get a reply good enough to keep me here as a reader. Otherwise a new tech site will be part of my daily tech edification.

The only "weakness" they reported was the hold up time lower than the standard of 17ms - the value measured was 13.8ms for 230v and 14.5 for 115v. It's on this page: http://www.tomshardware.com/reviews/gaming-psu-efficiency,2796-26.html

According to the ATX PSU Design Guide v2.01 (http://www.formfactors.org/developer/specs/ATX12V%20PSDG2.01.pdf), the hold up time is 17ms :

3.2.11. Voltage Hold-up Time
The power supply should maintain output regulation per Section 3.2.1 despite a loss of input power at the low-end nominal range—115 VAC / 47 Hz or 230 VAC / 47 Hz—at maximum continuous output load as applicable for a minimum of 17 ms.

So the Seasonic is disqualified because while powering 700-750 W of electronic equipment it was only able of keeping the system on for 14ms instead of 17. Ridiculous.

I'm positive at regular gaming PC load of 250-450W, the power supply would have no problems keeping the system powered on for much longer time... not that it would actually matter, in real world as any UPS would switch on in about 5-10ms and otherwise the system would power off anyway.

At the same time, the ripple outside the standard by far of the Cooler Master GX 750 seems to not be a reason to disquality the power supply:

http://www.tomshardware.com/reviews/gaming-psu-efficiency,2796-9.html

In the remaining tests, Cooler Master's GX 750 fully meets the expectations of an 80 PLUS power supply, even if its efficiency drops a bit under low loads. It does mess up a bit with the ripple and noise though, giving us a reading on the 3.3 V rail that is 50% above the ATX specification limit. The good results for the 5 and 12 V rails suggest that this is a weakness in the PSU design.


The Cooler Master GX 750 is the only basic 80 PLUS-certified unit in this roundup, and is faced with almost overwhelming competition. It holds up fairly well, though. Its modest price and performance are appropriate, though the small ripple problem overshadows our general impression, especially with so many other good units in our roundup.

Yeah, right. The 50% OVER the maximum permitted in the ATX standard is a "small ripple problem" (after all, ripple on 3.3v can only kill your RAM and your SATA drives esp. SSD drives) while a useless hold-up time that's only 10-15% less than the minimum is serious enough to disqualify the Seasonic model.


Another thing worth mentioning, for the Cooler Master Silent Pro Gold 700 psu...

http://www.tomshardware.com/reviews/gaming-psu-efficiency,2796-11.html

At the bottom it says:

The ripple and noise tests reveal no weaknesses; the PSU is well within the norm. All other test results are unremarkable as well. Even the noise level under high loads is good.

The Cooler Master Silent Pro Gold 700 is a PSU with very high electrical quality. It is well-made and stands out for its high efficiency.

But the graph between these two lines of text shows that for 5V and 3.3V, the ripple is 30mV and 40mV - the maximum permitted is 50mV.

So if this power supply is o "very high electrical quality" and "well within the norm", how would you rate the Seasonic with 20mV and 30mV.... ultra mega uber high electrical quality?

 

[reasonablevoice]

Shouldn't you mention that Cooler Master did all your tests for you? Because that is a HUGE conflict of interest, especially when their crap unit with crappy ripple got rated second best.

 

[neosoul]

[citation][nom]scook9[/nom]You are either lying or very unlucky (got a bad PSU)......I have crossfire 5870s and an i7 965 and all stock cannot exceed 650W at the wall (about 550W actually used) no matter what I try.[/citation]

PC Power and cooling are excellent PSU's. What I see happening is that he/she have an older model that didn't supply enough 6-pin connectors and you had to do some goofy molex splitters.

 

[anonymous x]

[citation][nom]p4l1ndr0m3[/nom]So, since I cannot read. Why don't you tell us all what test the Seasonic didn't pass? I quote directly from the article, "The bottom line is fairly positive with the Seasonic X-750 scoring great results in our efficiency tests. The supply is not without weaknesses, though. Seasonic is generous with its eight Molex and nine SATA connectors, as well as with the three CPU connectors." No mention of what those weaknesses were. Anyone who agrees with the findings in this article either do not know better or blindly support anything Tom's Hardware writes. I've written a final complaint/plea to Tom's in the Site Feedback/Website Opinions section. I am hoping to get a reply good enough to keep me here as a reader. Otherwise a new tech site will be part of my daily tech edification.[/citation]
The guy above me replied to you about it, but I'll post it again to make sure you read it as that is something you seem to have a difficult time doing.
http://www.tomshardware.com/reviews/gaming-psu-efficiency,2796-26.html
Ridiculous requirement or not, it is still part of the ATX specifications, still something that was tested, and makes the conclusion perfectly valid considering the criteria chosen for a recommendation.

 

[Guest]

Hahahahahaha.

Did you read your own article before making the above conclusions??

 

[Reynod]

Just finished reading the review ... not bad ... so I don't see what all of the fuss is about.

crash ... it seems you and the other guys didn't get much of a break at christmas ??

I guess you will be busy testing Sandy Bridge systems now ?

 

[i_am_aronman]

[citation][nom]santiagoanders[/nom]There are power supply fanbois now? Your beloved Seasonic wasn't recommended so you rant?[/citation]

No i think that he is saying that it did really well on the tests and is wondering why they did not even get a mention. I don't thank he was being a fanboy he was just stating the obvious to people that did not read the whole article. (Not saying you did not, everyone is entitled to their own opinions.)

 

[Guest]

To the several who said variations on "Why didn't you test ______?"

That's answered clearly on the first page, fools. "we asked PSU manufacturers to send us products developed for a very specific and very demanding group: gamers."

So they tested what the mfrs SENT. Is it that hard to understand?

 

[bildo123]

I've hate to be biased on a one-time-basis as it's unfair usually, but I've had a really bad experience with getting a Seasonic X650. I believe the PSU itself worked well, but there was this constant irritating buzzing whenever there was a low load. When I played games it went away for the most part. But if I was just sitting there with Word open typing a report, the noise was very apparent. I read some reviews and found this to be an issue with other X750/650 PSU and decided to possibly RMA it. I believe I called at least twice and emailed three different times and never-ever-at-all got any type of communication established over the span of two weeks. At that point I just dealt with the seller for a full refund which got me whacked for shipping. Not impressed. Again, it was probably just a lemon, but I think the whole non-existent customer service deal just turned me off completely.

 

[jdamon113]

I have been reading Toms for a long time.
And this test to sound like they got paid for this comparison.
First the Antec is the same psu an the seasonic, seasonic makes this item for antec. As they do in many of there lines.
The Nzxt is hiding there uL number wiht is a sign of warrning. I did some looking it is make by TopPower, a related Toms artical a while ago suggest some of these no name brands are not recomended.
The Chieftec is the worst in the bonch not sure how this is recomended at all.
The cooler Master SP is made by AcBel, they are a oem builder os psu and generally have a good qc and some saftey devices in the psu.
But not up the the quility of seasonic, emermax and Startech...

What gives guys, did you get paid for this one.
for the record, I would consider the antec, but the other three I would not take if you gave them to me.

I would say do a real test on the psu. Tru amprage and how about life cycle and build quiltity.

Look at the best rated brands. YOu will find seasonic enermax, corsair which is seasonic, antec with is seasonic, and toppower. PC and powercooling some are seasonic and some are Win-Tach and have a very high build quility.

Bad artical Tom, you should look at previous articals.

 

[SabreWulf69]

I still reckon the best PSU is the 1250W Enermax Revolution 85+. On this test though, the Corsair AX850 230/240V version seems to have checked all the right boxes, what about that one?

 

[studioman22]

Yeah well speaking of holdup time, I don't understand why these PS manufacturers haven't included basic power loss protection for like 2 to 3 seconds. A very brief power outage is what happens most of the time, and then the power comes right back- but it's often enough to shut the system down. Instead we have to buy an expensive online UPS. There has to be a way to do this on board the PS, even if the PS costs more.

 

[mariushm]

Yeah well speaking of holdup time, I don't understand why these PS manufacturers haven't included basic power loss protection for like 2 to 3 seconds. A very brief power outage is what happens most of the time, and then the power comes right back- but it's often enough to shut the system down. Instead we have to buy an expensive online UPS. There has to be a way to do this on board the PS, even if the PS costs more.

The answer is simple... have a look at the insides of the Seasonic X-650:

See those round things, white with a blue band, and those small round brown? They're called capacitors... to be able to keep the voltages stable and steady for even a second at 650w, these capacitors would probably have to be as big as those two capacitors on the input side - there's simply no room in the power supply for such big capacitors and even if you'd somehow make them sit over some integrated circuits, the air circulation inside the case would be horrible, components would heat and so on.

You could replace them with ultracapacitors which MAY be able to provide a few seconds of time in about the same diameter (but they'd be higher) but at about 3-5$ a piece versus 10-30 us cents for these regular capacitors, you're basically going to double the cost of the power supply.

Besides, I'll have to disagree with you. Where I live, there's power failures about 3-4 times a year... maybe once for less than 10 seconds and the other 2-3 times for about 10-15 minutes and the whole neighborhood is offline.

To power a computer these power supplies are designed for (about 300-400w constant usage when playing games) an average UPS that costs about 90$ has a battery that weighs about 6 Kg, it can keep the system running for 4 minutes and then needs to recharge the battery for 16 hours until it can keep up the system running again at that power usage level.

Even if you'd make a UPS that would hold the power for 20-30 seconds it would still be about 2 Kg, it would be just as big as a power supply, it would need about 5 hours to recharge after a power loss (during which it would be useless), those 2kg extra would probably increase the shipping price of the psu by about 15-20$, and so on and so forth...

 

[sblantipodi]

this reviews is good BUT I need to know at what load they have tested the hold-up time.
please help.

 

[Guest]

No PC Power & Cooling products?

 

[sasikanth8]

Update with new psu's please

 

[pspd96]

It would have been nice if you included XFX they really make good power supplies