POST A COMMENT

41 Comments

Back to Article

  • Stuka87 - Monday, March 26, 2012 - link

    On the last page where things are being highlighted, shouldn't the Corsair be highlighted for efficiency and ripple since it was ahead in these categories? Unless the yellow is just meant to show there is a difference. But it seams in the rest of the areas the better supply is highlighted.

    Both supplies look pretty good though, even if priced a bit higher than some similar supplies.
    Reply
  • Martin Kaffei - Monday, March 26, 2012 - link

    Sounds more correct, right.
    Thanks a lot.
    Reply
  • lyeoh - Monday, March 26, 2012 - link

    Has Corsair really improved their quality since: http://www.behardware.com/articles/843-3/component...

    FWIW, OCZ and Corsair's nonPSU stuff seem to get returned quite often too:
    http://www.behardware.com/articles/843-4/component...
    http://www.behardware.com/articles/843-7/component...
    http://www.behardware.com/articles/843-8/component...

    Both these PSUs might be OK, but why should I care? There are other manufacturers who have a better track record when it comes to quality. Poor quality in a range of similar models is understandable, but poor quality in different types/classes of products just tells me to avoid buying anything from them.

    I'd prefer to see more reviews of products from manufacturers that care about quality. Who the heck wants to read reviews of food from restaurants that have a long track record of giving a significant percentage of their customers food poisoning? I don't care how good their food looks.

    Are these even significantly cheaper?
    Reply
  • Samus - Tuesday, March 27, 2012 - link

    My advice to you is stop reading BEHardware.

    Those statistics for every segment they 'measured' are a joke. Just drafting such a distasteful article is a huge lapse in journalism judgement on their part. Nobody in their right mind would gauge component reliability based on return rates over a 6-month period from a single source.

    The only way to measure reliability is by reviewing the initial quality in detail and putting the components into real-world scenarios to report any abnormalities over time. AnandTech and other legitimate hardware review sites do this, as does Car & Driver, Consumer Reports, Home Theater Review, Gun Directory, etc
    Reply
  • seanleeforever - Tuesday, March 27, 2012 - link

    while i agree with your point of view, but AnandTech rarely review abnormalities over time. and bear in mind AnandTech is more akin to C&D not CR because AnandTech receives cherry picked product directly from manufactures. and you really have to hope what you buy from your local frys is the same as the one they choose to send for reviewing. Reply
  • lyeoh - Thursday, March 29, 2012 - link

    6 month period? OCZ's poor track record goes way back longer than that: http://www.behardware.com/articles/831-7/component...
    http://www.behardware.com/articles/831-4/component...
    http://www.behardware.com/articles/810-6/component...
    http://www.behardware.com/articles/810-4/component...

    You haven't even provided a good reason why anyone should care about your opinion on this subject especially since you said "6 month period".

    I'd believe BeHardware's article more than I'd believe you, especially since a few simple Google searches for product problems do not disagree with their statistics. So it's not just a freak incident where only the french retailer has higher rate of returns for OCZ and corsair stuff.
    Reply
  • fausto412 - Monday, March 26, 2012 - link

    I have a 6990 and a n AX850 PSU...that psu gets super hot when gaming. Can't imagine a 550 watt psu not melting. I think the days of the 550 watt psu are over unless you only do single card gaming and don't use the highest end cards. Reply
  • Ken g6 - Monday, March 26, 2012 - link

    Did you miss the GTX 680 review the other day? http://www.anandtech.com/show/5699/nvidia-geforce-... The 7970 would work fine with a 550W PSU too.

    Also recall that the 6990 is effectively two cards in crossfire on a single board.
    Reply
  • Samus - Tuesday, March 27, 2012 - link

    I run a GTX570, Core i7-950, 12GB RAM, 5 hard drives and some other shit off a PCP&C 500-watt PSU. Whole system is reasonably warm after hours of gaming and is completely stable.

    Kill-A-Watt has never reported more than 400-watts drawn from the wall, so my 500-watt PSU never even hits 80% load.

    Any never system with a single GPU draws less power, I guarantee it. 550-watt is overkill for most people with current technology. Remember most OEM systems still ship with 250-watt PSU's.
    Reply
  • Stuka87 - Monday, March 26, 2012 - link

    I use a Antec NEO-ECO 520C with my over clocked 7950 and its perfectly fine. I never get anywhere close to even 400W at peak (And thats with a 4.1GHz Phenom II) when measured at the wall. The latest chips from AMD and nVidia are incredibly efficient. The 6990 is a pretty big power hog comparatively, and it is a dual GPU card. Reply
  • Rookierookie - Monday, March 26, 2012 - link

    >I think the days of the 550 watt psu are over unless you only do single card gaming and don't use the highest end cards.

    That's like saying that the days of the graphics card is over unless you play games and turn on graphic features.

    Other people do exist, you know.
    Reply
  • retrospooty - Monday, March 26, 2012 - link

    " I think the days of the 550 watt psu are over unless you only do single card gaming and don't use the highest end cards."

    Off all gamers, PC gamers are the minority.
    Of PC gamers, high end cards are the minority
    Of high end cards, dual chip/SLI is the minority

    You are talking about the minority, of a minority within a minority.

    Of course this is a tech site and more of us here are likely to game on PC and have extreme setups and overclock to the highest possibility, but we are just a tiny fraction of the market.
    Reply
  • bji - Monday, March 26, 2012 - link

    Not to dispute your point, but this is a PC component site so non-PC gamers are not reading these articles. Therefore your first categorization is irrelevant.

    It is true though that high end cards are in the minority in PC gaming, and SLI a minority within the high end card market. So the fraction we're talking about is still small, just not as small within the audience of Anandtech articles as you described.

    To try to stay on point: one of these power supplies comes with a feature that would prevent me from buying it, no matter what: the OCZ badge on the case.
    Reply
  • gamoniac - Monday, March 26, 2012 - link

    My OCZ Fatal1ty 550W's 4+4 pin was only 45cm, posing an issue reaching the farthest 12v ATX connector if I route it behind the motherboard (I had to get an extension). It looks like OCS has listened and increased the length to 60cm. Well done.

    Corsair TX seems like a good choice, but I am not sure why they need to include two FDD connectors to clutter up the case, especially if it is not a modular PSU to begin with. Does anyone still have an FDD on their PC, let alone two? I got rid of my FDD a few years ago after collecting dust in my garage for a few years.
    Reply
  • HowQuaint - Monday, March 26, 2012 - link

    Could you be more specific about the "magnetic amplifier" you specify in the OCZ unit? Wikipedia has an article about it but it doesn't seem to be a power supply technology. Since mag amps use a transformer-like device I'd guess it's a forward or flyback converter, which is more commonly used for the primary, high-voltage stage but it still works for low voltages. But saying that it "has no DC-DC converters inside" is inaccurate because any sort of secondary step-down qualifies as a DC-DC converter regardless of topology, and with cross-load numbers that low it definitely uses a secondary converter. Reply
  • Martin Kaffei - Monday, March 26, 2012 - link

    Wikipedia is probably not the best source for information like this. As you might see there are two chokes with a black core between the 5VSB and main transformer. Those are the chokes for regulation directly behind the windings. During saturation the voltage comes through. During the (let's call it) "high impedance mode" a part of the voltage pulse gets cut.

    I didn't say there are no DC-to-DC converter. I know it is the most famous an easiest converter type. But in this case I'm talking about step down converter for 3.3V and 5V, who get current from 12V. Each output voltage has it's own regulation. Most common PSUs regulate 5V and 12V with the same IC which is bad for crossload regulation.

    Hope this was understandable since English is just my second language.
    Reply
  • Leyawiin - Monday, March 26, 2012 - link

    The SeaSonic M12II 620 Bronze 620W is the exact same price as those two at Newegg and is a more robust and reliable power supply. I wouldn't touch one made by Great Wall or CWT when a superior SeaSonic built unit is available at that price, especially with a higher capacity. Reply
  • dubyadubya - Monday, March 26, 2012 - link

    +1 good catch I agree 100%. That Seasonic is a way better deal, Better quality, more power for the same price its a no brainer. I'd own a CWT PSU but its a step down from a Seasonic brand or built by Seasonic PSU for sure. Personally never seen a Great wall built PSU as far as I know. Reply
  • johan851 - Monday, March 26, 2012 - link

    Definitely. Pointing that out made me drop consideration for these - Seasonic makes a much better supply. Reply
  • TegiriNenashi - Monday, March 26, 2012 - link

    Objective noise measurement is easy: there is free sound meter programs both for iPad and android. Reply
  • A5 - Tuesday, March 27, 2012 - link

    I wouldn't really call an SPL Meter app coupled with a bad tablet/phone microphone "objective". To get true, repeatable SPL numbers, you need to have a special test setup like SPCR. Reply
  • MeanBruce - Monday, March 26, 2012 - link

    And the Corsair even looks better. Ok ok we all love Corsair PSUs, so is it now standard protocol for the company/groups on the bottom to offer new technology first and the group on top of the market to offer the same technology LAST? Am I the only enthusiast wondering where the Corsair Platinum PSU line-up is hiding? Is the strategy to wait and see what all other groups produce, then go to the drawing board simply to one-up the competition and stay on top of the market? Reply
  • johan851 - Monday, March 26, 2012 - link

    One nice thing about not gluing down the fan connectors is that it's easy to replace that cheapo fan if/when it fails. :) Reply
  • jabber - Wednesday, March 28, 2012 - link

    Yeah I finally got round to replacing the horrendous fan in my Corsair TX750 MK1.

    Sounded like a leaf blower after about 15 minutes of use. Dont quite know why it had to blow at 3000rpm all the while.

    Swapped it out for a low start power 1700rpm fan and its been a lot better.
    Reply
  • Rick83 - Monday, March 26, 2012 - link

    I'm not really convinced that you can measure efficiency anywhere near a hundredth of a percent, accuracy wise, never mind that at that scale the slightest of atmospheric changes, input voltage changes, inter-series variation will void those results instantly.

    Especially using variations in those non-significant digits to mark advantages of one solution over another is a bit misleading. So unless you're absolutely convinced of those numbers, you should stick to values that you can reasonably expect to hold up under scrutiny.

    Especially if your sample rates are 1, you should assume a certain variation, and ideally if over the course of testing you measure an average, also report the average std-dev.

    Benchmarking is an engineering discipline!
    Reply
  • ssj3gohan - Wednesday, March 28, 2012 - link

    Wow, I never thought I'd meet somebody who I would agree with so much on this subject. Yes, measuring and properly representing your measurements is a key, key factor in benchmarking. Anandtech usually has their methods straight in other areas (although they consistently fail to include statistical parameters, really people, just because other sites don't show it doesn't mean you are allowed to omit them!) but power supplies are again (like I said in a lot of earlier PSU articles) their bastard child. Only results in table format, bad result representation in general, no methodology, no ambient conditions mentioned or tested against, very limited and unimaginative testing for that matter... it's like the exact opposite of, let's say, the SSD articles. As my go-to tech review site, this is extremely disappointing.

    Of course, I'm totally biased for being a PSU designer myself.
    Reply
  • Chris Simmo - Monday, March 26, 2012 - link

    I have been using GW PSU's for years (OE) and while they may not be quite as efficient as is stated here, they are incredibly reliable. We have used approximately 3000 PSU's in the PC building company we work for and had less than 20 fail over 4 years. The Corsair PSU's have been rubbish and we refuse to sell them. I have had customers bring us parts for assembly. Of 5 PSU's, 3 were DOA (TX750, TX850, AX850). I don't give a rats arse how efficient they are, they are not worth the money with this horrid a failure rate! Seasonic make awesome PSU's too, but they cost a pretty penny in comparison to the GW PSU's. Good on OCZ for picking a reliable company, not just for efficiency Reply
  • Beenthere - Monday, March 26, 2012 - link

    Both OCZ and Corsair have a sketchy history with PSUs in addition to other products. I would recommend buying a Seasonic or PC Power and Cooling PSU if people want performance and reliability. Reply
  • Holly - Tuesday, March 27, 2012 - link

    Page 6, first sentence says "Corsair uses a 140mm fan from Yate Loon with a ball bearing." ... shouldn't it be "OCZ uses..." ?

    And as for Seasonic PSUs... I have had 6 of those in my hands last year in time span of 6 months. Out of those 3 had to be replaced due to heavy instability (one caused frequent BSODs, one caused system restarts and one air fan was extremely loud) immediately, two other gradually caused high system temperatures and the last one simply didn't start up after about 2 months usage. They were calculated so the system wouldn't load them over 60% at peaks.

    It might be just a bad luck or Seasonic commits low quality products in my country, but I would never recommend or buy them again. I have had variety of Corsair PSUs in my hands from CX, TX, HX, VX and AX line, had to replace only one CX 500W because of excessive fan loudness.
    Reply
  • Martin Kaffei - Tuesday, March 27, 2012 - link

    You have good eyesight. Fixed and thanks. Reply
  • C'DaleRider - Tuesday, March 27, 2012 - link

    Why do you persist in posting needlessly obtuse and difficult to decipher voltage regulation charts?

    Instead of making me do the math....such as figuring out what the +3.3V rail is at +0.61%, why don't you instead simply post what the damned rail started at and drifted down to under loads, such as +3.4V, +3.35V, +3.3V, etc.

    No...you have to try to be "different", and by being different you have become obtuse and difficult to decipher.

    Simple is good, complex is needless and just silly. I'd suggest you look at the way JonnyGuru and HardOCP lists their voltage and ripple/noise charts to get an idea of what nice, clean, and simple presentations of the information can be instead of your almost useless charts.

    When I read reviews, I don't want to be bothered with having to do the calculations in my head to figure out what you mean by a +1% of the +12V rail means. (I'm guessing it's +12.12V, right? That's not hard, but then you get to +0.62% or -0.24% and that becomes stupidly hard.)

    Sometimes trying to be different isn't better, just silly.
    Reply
  • bobsmith1492 - Tuesday, March 27, 2012 - link

    Percent change is a better metric especially comparing across different voltage rails. Everything electronic is specified in percentage tolerance. So percentage measured is the obvious choice.

    1% of 12V is pretty easy if you want absolute values for some reason. 12x1.01. I suggest you go back to middle school.
    Reply
  • philipma1957 - Wednesday, March 28, 2012 - link

    It arrived yesterday. Newegg had it for 69 on shellshocker 20 rebate is 49 and a 15 dollar gift card my cost was 37 dollars due to sales tax. At 37 bucks it is a good deal. I think I will put in a better fan. I will look to see if I can get the sanyo denki fan that the author has picked out. I will also add a piece of insulation between the modular pcb and the main pcb.

    Reading this info has inspired me to check on my 660watt seasonic psu as every once in a while the computer shuts down for 1 spilt second, it will not hurt to take a look. To me the biggest problem with the ocz in the review is the lake of input voltage regulation. This means a good ups/voltage regulator should supply the psu. I have that. I have some good thin insulation. So i will need to get a fan. well so far only 37 spent on this psu. I wonder what the sanyo fan costs.
    Reply
  • Beenthere - Wednesday, March 28, 2012 - link

    The Sanyo fan is $45.99... :( Reply
  • philipma1957 - Friday, March 30, 2012 - link

    Thanks for the fan price. Too much so I grabbed an antec true quiet 140mm fan for about 15.50 at amazon. Both parts arrived on the 29th. I tested it with the stock fan and the antec a lot quieter. So my 37 dollar bargain is now about 52 bucks. I have a lowcost Nxzt case and I was able to route the antec's fan switch behind and lower then the cpu on the other side of the case. Since the case has easy access to that side panel I can switch the fan to low setting if I choose. I also was able to attach the fan's power cord to the mobo So I can monitor its speed .

    The unit appears well built and the solder work inside the case looked clean. I added a thin piece of rubber as suggested to prevent any arcing issues between the power jack pcb and the main pcb. I will get back to this with an update to let you know how well the unit holds up in daily use. I paid a total of 52-53 for this.

    my 660 watt seasonic was 116. Still not sure if the seasonic has an issue or some other problem cause rare spilt second black outs every once in a while.

    this ocz is going into my amd lower cost build. about (500)

    while the seasonic is in my intel higher cost build. about (1k)
    Reply
  • philipma1957 - Monday, April 02, 2012 - link

    Amd fx-4100

    2 x 4gb patriot/amd 1333 ram

    gigabyte 970 atx mobo

    a crucial m4 128gb ssd

    a coolermaster 212 evo

    a sapphire hd6670 passive cooled gpu

    The psu runs very quiet . the antec fan is much quieter the the stock fan. seems like a good buy.
    Reply
  • StrangerGuy - Monday, April 02, 2012 - link

    $90 for a 550W PSU? Are you kidding me? And it's ONLY 80+ Bronze. I got a Seasonic S12 520W for only $60.

    I can see why there are so many players in the PSU market, all thanks to the enormous profits to the PSY FUD crowd who spouting nonsense like "400W isn't enough for the average gaming PC" brings.
    Reply
  • philipma1957 - Tuesday, April 03, 2012 - link

    you are corect it is not worth 90 bucks. but AR it is 70 at amazon. I paid 50 AR and I a got a 15 dollar gift card to boot. so it was 35 bucks. the fan was to loud and with the mod I did it was 52 bucks. it is worth 52 bucks. Reply
  • mikeymikec - Thursday, April 05, 2012 - link

    Considering how a desktop PC without a graphics card and a decent PSU can run virtually silent these days, I think more work could have been done in the audible noise from PSUs.

    My first thought in how to achieve this would be to do a review of 12-14cm chassis fans, pick the most quiet one, then use that as a reference for PSU testing in future. Give an opinion on the noise level on each PSU reviewed like you've done here (because fans can emit extra irritating noises like a whine while still being quiet, but also compare it to that one, also possibly measuring the noise level and providing measurements of, say, +2dB (or whatever is considered the best method of measuring audible noise levels) above reference fan.

    I've bought a lot of Corsair VX range PSUs and a few since the VX range was discontinued, and the TX ... M range are much noisier, as in, they're easily the highest noise producer in a modern system. At least with the VX range the noise level was approximate the same as the CPU and chassis fans in the system (without either of those being unusually decent fans, a stock CPU fan and a Coolermaster 12cm fan that comes with the Coolermaster cases I typically use).
    Reply
  • mikeymikec - Thursday, April 05, 2012 - link

    First para: I meant, more work could have been done to measure/analyse/give a useful opinion on the audible noise from these PSUs. What I can take away from this review is "they produce some audible noise". Reply
  • mikeymikec - Thursday, April 05, 2012 - link

    Argh, it really goes to show how much I rely on the 'edit' button... PROOF READ BEFORE POSTING! :) Reply

Log in

Don't have an account? Sign up now