Want to read Slashdot from your mobile device? Point it at m.slashdot.org and keep reading!

 



Forgot your password?
typodupeerror
×
Hardware

Gigabyte Will Replace 'Exploding' PSUs, Downplays Risk (kotaku.com) 72

UnknowingFool writes: Gigabyte announced it will replace PSU models GP-P850GM and GP-P750GM that fail after HardwareBusters and Gamer's Nexus both showed their PSUs fail catastrophically in small explosions. Gigabyte however downplayed the risk by misstating the failed PSUs only exploded after prolonged extreme conditions and that the problem had been corrected last year. Both HardwareBusters and GamerNexus contest multiple points from Gigabyte. Both stated that their units failed quickly under moderate to light loads. In the case of Gamer's Nexus they obtained some units as late as May 2021.

This issue was first reported in November 2020, when Hardware Busters evaluated a GP-P750GM model and found it exploded during testing. After many Newegg customers reported multiple issues with two Gigabyte models, Gamer's Nexus purchased or obtained units from customers over the course of several months in the spring. Their testing found 5 out of 10 units failed quickly during OPP (Over Power Protection) testing. A normal OPP scenario should have had the PSU shut down to protect itself. Instead, some of the units exploded.

A further complication arose when Newegg customers tried to return failed units. Many of the PSUs were bought in Newegg Shuffle bundles with GPUs, and customers would have to return both the GPU and the PSU. Considering the GPU shortage, most Newegg customers complained about this requirement.

This discussion has been archived. No new comments can be posted.

Gigabyte Will Replace 'Exploding' PSUs, Downplays Risk

Comments Filter:
  • Collateral damage?

    • Re: (Score:3, Informative)

      by Chas ( 5144 )

      Can't prove it.

      • by 2TecTom ( 311314 )

        Can't prove it.

        we already have a smoking power supply what more do we need? Clearly, this should be a class action, and the companies need to be held to account, just as a person would be ... sadly, i bet this just becomes lawyer fodder.

        • Class action? Enjoy your shiny nickel as compensation.

          • It would still be pretty massively expensive for Gigabyte to deal with a class action lawsuit, plus most of the wider tech press would pick up and run with the story further damaging the brand.

            If the warranty with the unit still includes the mandatory arbitration to settle disputes people could launch an arbitration bomb like they did against Uber and some other places. Usually the company agrees to pay all the fees associated with the arbitration so, you get a few thousand people all filing individual arbi

        • There should be no class action, leave the lawyers out of it and instead a recall that replaces the units for free would cost everyone less money.

          -Or do you just want a 5 dollar coupon while paying for the lawyer's next yacht?
        • by Chas ( 5144 )

          The problem is, yes, we have a kablooied PSU.

          But the PSU is supposed to have protections built in.

          So Gigabyte's gonna lean on that and go "Wasn't me..."

    • Given how Gigabyte is downplaying the current situation, I do not have high hopes that Gigabyte will reimburse customers. Complicating this is that it is difficult to prove that the PSU caused collateral damage. Gamer's Nexus lost a RTX 3090 and though it coincided when the PSU failed, they cannot prove that the PSU failure caused the GPU failure.
      • Doesn't the US has some sort of customer protection organism in place? Because besides collateral damage, a PSU which sparks and shorts is a clear fire hazard.

        • Well organizations like UL and CE although they're not consumer protection agencies in the sense meant, but their influence is important.

          Also one can wonder if this situation is a symptom of the global parts shortage?

          • by Junta ( 36770 )

            I haven't seen a specific analysis to really say one way or another, but reports that previously unknown component suppliers were showing up in them may indicate a manufacturer desperate to keep going lowering their supplier management discipline to keep making product.

        • The US does have an organization; however, consumer protections are weaker than in the EU. Many measures are driven by public outrage more than strong regulations. Unless an explosion injured a baby, a puppy, and a kitten simultaneously, do not expect much regulatory pressure.
      • Interesting since some battery manufacturers have a "if our product damages your property" in their warranty.

        • by Junta ( 36770 )

          If it actually catches fire and burns things, that's easy to prove. If a computer component failed due to an out-of-spec input voltage or current, then it's generally difficult to actually prove that happened versus some other failure that causes the device not to work. If there are scorch marks on the power pins, you have some good evidence, but generally a failure is more subtle, the card *looks* fine and you need some professional failure analysis to determined the precise cause.

      • "Cannot prove"

        GPU was fine before.
        PSU exploded into a smoking crater.
        GPU no longer works with new PSU

        What is there to prove? It's obvious, and bad customer service to not make things right.

        • Gigabyte: "Correlation is not causation. LALALALALALA. Can't hear you."
        • > What is there to prove?

          GPU developed a fault that overloaded the PSU, causing it to release the magic smoke.

          Prove otherwise.
          =Smidge=

        • by gweihir ( 88907 )

          "Cannot prove"

          GPU was fine before.
          PSU exploded into a smoking crater.
          GPU no longer works with new PSU

          What is there to prove? It's obvious, and bad customer service to not make things right.

          Simple: That the GPU was well designed. It looks to me that it was not, because a regular PSU failure does not result in overvoltage at the outputs and hence the GFX card should not die. A badly designed GPU voltage regulator may overload in too low voltages, but that is pretty much the fault of the GFX card design.

        • Do you have any witnesses?
      • by gweihir ( 88907 )

        If they lost an RTX 3090, that is likely due to poor design of that card.

      • The PSU that the 3090 failed on never died, and the computer was still running, the GPU had just died. Thus why they say it may have just been a faulty 3090(which was a Gigabyte model anyway making them still at fault).

    • by gweihir ( 88907 )

      Collateral damage?

      Unlikely. Well, maybe some lost documents if they were open. This is not an "explosion" in the conventional sense, rather a controlled "burn off". Electronic components are designed with failure and overload in mind. Also, due to the design of these PSUs, component failure typically does not result in overvoltage at the output. That is for classical transformers when they short out in some specific way.

      Still should not happen and Gigabyte clearly cut safety-margins down way too much. You know, like an ElChe

  • Don't care what the price is or how Gigabyte protests their innocence.

    These PSUs are crap product. Gigabyte and Newegg KNOWINGLY pushed this e-waste on people with the Newegg Shuffle and then tried tying WARRANTY SERVICE to the video card as well. So you have to return the video card if you blow the PSU.

    Screw Gigabyte.

    • So you suggest to buy Samsung PSUs ?!?
  • Exploding PSUs is natures way of saying, step away from the computer, go outside and get some fresh air.

    • Exploding PSUs is natures way of saying, step away from the computer, go outside and get some fresh air.

      But it's scary out there! They'd have to deal with people at some point.
  • by Lisandro ( 799651 ) on Tuesday August 17, 2021 @11:17AM (#61701239)

    ...is pretty damning. They're making PSUs go boom with 5 minutes of heavy load. [youtube.com]

    Gigabyte is quite literally playing with fire, IMHO. Just recall the damn things.

  • by cloud.pt ( 3412475 ) on Tuesday August 17, 2021 @11:19AM (#61701251)

    I've been following this topic for months now. It's not only amazing the time it took to Gigabyte to address the issue (pretty much waiting it out for a large influencer such as GN to chime in...), but most importantly, it is absolutely disgusting the way retailers, likely by Gigabyte's influence itself, have been pushing these faulty, dangerous units to the market through bundles and discounts that are NOT REFUNDABLE under acceptable conditions, pretty much leaving the consumer to hold the bag.

    Even their latest action is pretty much to shift issues to consumer - the most affected party as they may actually be harmed physically or their property by use of these products. Gigabyte is NOT developing these products, just branding them, and stating that reviewers aren't placing "normal use" on these while at the same time citing meaningless, self-validated regulatory certification just as an excuse that the product is safe... It's absolutely irresponsible.

    I for one can see a lot of evil in Gigabyte, part of the huge Foxconn, in flipping these just for their own financial protection. Just the fact they decided to release information about their supposed hack THE VERY DAY GN PUBLISHED THEIR FINDINGS is shady enough for me to never consider their products ever again at least as a top choice or anywhere near it. This is clearly a company who is showing total disregard of consumer safety in favour of keeping this under wraps so they can keep selling the unit incospicuously to consumers.

    Just take note of the amount of fuckups around this PSU line:
    - used different fan bearings than advertised
    - use absolutely horrible FETs that blow up after OPP trips, not before
    - uses different parts in the same models which have very different behavior
    - use unbranded parts that not even the best PSU reviewer around (Aris) has never seen before
    - kept this under wraps while flipping units for discounts
    - aren't accepting full refunds to owners who may have had to purchase something else to avoid getting a fire in their houses
    - and worst of all: they silently, **and only partially fixed** the issue on undisclosed serialized products, with a mechanism that Aris and GN themselves consider isn't the root cause of the problem...

    What a clusterfuck.

    • by Junta ( 36770 )

      - uses different parts in the same models which have very different behavior

      This was the one part of the GN video that I think detracted from their point. At least in the video, they didn't point out different behaviors in those specific components, but instead generally criticized having multiple sources for commodity components. Having multiple sources for a given capacitor or voltage regulator is not, in and of itself, such a horrible thing. This is generally accepted as the correct strategy, to have 2 or 3 possible vendors for any given component whenever possible. This is gr

      • This was the one part of the GN video that I think detracted from their point. At least in the video, they didn't point out different behaviors in those specific components, but instead generally criticized having multiple sources for commodity components. Having multiple sources for a given capacitor or voltage regulator is not, in and of itself, such a horrible thing.

        But that was not the totality of that criticism. Part of the criticism was that they did not know the companies used. And Gamer's Nexus has extensive knowledge of these vendors. The main point was not that a company cannot multisource parts but they should be sure that all sources are equally good. When a product uses parts from unknown vendors, that is a cause for concern.

    • by AmiMoJo ( 196126 )

      Just remember they can only do this because you have extremely weak consumer laws that allow them to. They aren't doing it Europe because they will get the book thrown at them if they even try.

      • by nnull ( 1148259 )

        We are supposed to have strong protection from this in the form of third party certifications, such as UL. Lately, they have been holy water blessing everything or manufacturers making faking certifications, so you see this garbage come through. The same thing is happening with solar stuff, everything has been holy water blessed, while house fire after fire, nobody cares.

        Or my favorite thing manufacturers do, they provide a CE self-certification in the US, which is absolutely meaningless here because we don

        • by AmiMoJo ( 196126 )

          Have you ever done UL? It probably wouldn't pick this kind of thing up. First you have to supply them with samples and you are then allowed to replace components with "equivalents" if you think they won't affect the certification. They generally don't do the kind of tests that GN did to make these PSUs fail either, they are more interested in emissions and the like.

          In any case the issue is not the certifications, it's that even though they know these PSUs catch fire they are trying to shift them by forcing

          • by tlhIngan ( 30335 )

            Have you ever done UL? It probably wouldn't pick this kind of thing up. First you have to supply them with samples and you are then allowed to replace components with "equivalents" if you think they won't affect the certification. They generally don't do the kind of tests that GN did to make these PSUs fail either, they are more interested in emissions and the like.

            UL is a laboratory. They do safety certification tests based on a standard.

            if there is a standard, then you can get it certified - that's why yo

            • by AmiMoJo ( 196126 )

              The most they would test is the over current protection, which works at the test parameters they use. The problem is that the shutdown threshold has been set way too high and if reached there is a good chance of an explosion. UL don't normally test until something destructively fails, at least for PSUs.

        • by gweihir ( 88907 )

          You are mistaken. UL mainly certifies safety, for example that this PSU does not burst into flames when overloaded. And it does not. It has a short, controlled burn-off and that is likely completely fine with UL. UL generally does not certify reliability or fitness for purpose.

      • by gweihir ( 88907 )

        Just remember they can only do this because you have extremely weak consumer laws that allow them to. They aren't doing it Europe because they will get the book thrown at them if they even try.

        Indeed. The whole thing sounds like a management fuck-up to me that the cretins responsible tried to cover up. In Europe that would not work.

      • Only they ARE doing it in Europe. The 1000W version (which hasn't been tested but is from the same manufacturer and which I already pointed to both GN and Aris for feedback) has been selling for for 120euro: a very discounted price from the original ~150 and under 40 of the standard 160 euro other 1000W 80+ Gold models go such as from Corsair or Seasonic.

        There aren't as many protections as most US folk would think. We have distance sales and standard 2y warranty, maybe some added regulation on retailer duty

        • Note this is also happening across the 750W and 850W models in Europe. Go to tropicalprice.com, look for the two models, and you'll see in every european Amazon they are not only available and heavily discounted compared to other Gold models, but a lot of them have bad reviews which report broken motehrboards, GPUs, hard discs, SSDs, USB ports on case, DoAs... And this product is still being sold. Amazon hasn't removed them for the issues.

  • I bought a PC that I configured at CyberPower during the gpu/cpu shortage, but I went with a Thermaltake Toughpower GF1 850W 80+ Gold instead. I couldn't choose the brand or model of RTX 3080 I got, but funny enough I received a Gigabyte one, the Gigabyte RTX 3080 Gaming OC 10G. It's been solid so far, and it has good reviews. I read up on the PSU reviews, and the ones for the Gigabyte PSU made it sound less than premium. Good thing I checked!
  • by Rick Schumann ( 4662797 ) on Tuesday August 17, 2021 @12:09PM (#61701509) Journal
    Back in the day when I was repairing arcade games for a living (such as it was, LOL) cheap shitty switching power supplies were just starting to power many of the video games. They were very primitive compared to a modern PC power supply, just the basics internally, and they could and would blow up just like in the youtube video linked in one of the TFAs: main switching transistors blown to Kingdom Come, which yes would make quite a bang and more than a little smoke. I'm shocked that in 2021 any company producing a switching power supply, with all the protections built into it, could be so flawed.
  • Gigabyte (Score:3, Informative)

    by Tailhook ( 98486 ) on Tuesday August 17, 2021 @01:19PM (#61701801)

    Gigabyte has always been on the ghetto end of the DIY components spectrum. The teardowns of these PSUs reveal the frenetic component sourcing going on; they source from whomever is selling the cheapest power mosfets and capacitors that day. They can't be doing more than cursory QA on that stuff.

    In the low margin, fungible world of PC components you get exactly what you pay for. Vote with your wallet; buy from the better component manufacturers. Good manufacturers, like Sea Sonic, offer 10+ year warranties and their PSUs don't blow up after 6 months. A 750W PSU from Sea Sonic is $18 more than the Gigabyte equivalent. There is a reason for that.

    • by gweihir ( 88907 )

      I agree on SeaSonic. Very solid, generous margins for overload, long-term stable, pretty silent. Not cheap, but that PSU will likely power more than one PC in its lifetime.

      • Fundamentally disagree, because I am at this moment shopping to replace a Seasonic 750 Focus Plus Gold that I bought 4 years ago based on Aris Mpitziopoulos' review/recommendation. It was very expensive, but it turns out that it does not actually put out it's rated power output, so it is completely useless in my new build, and will instantly crash and reboot when running any 3D application. Screw Seasonic. https://linustechtips.com/topi... [linustechtips.com]
    • yeah... but not quite.

      Note Gigabyte is pretty much Foxconn: they make iPhones, and they make some of the most reliable motherboards and GPUs. The problem here is they did NOT make these PSUs: they ordered them from another vendor who made them to some whatever spec, probably machined the exterior shroud and tested them... or one would hope they did LOL.

      The testing part, and most importantly the staunch defense of that testing is really the issue here: it was proven definitely NOT GOOD, yet they keep saying

      • I would say my experience with Gigabyte motherboards is awful as well. I build a LOT of machines for friends, family etc and I won't use a gigabyte anything anymore, gigabyte motherboards have been the one part that seems to fail early, I have never used their PSU's but now very glad I haven't
  • Exploded? Who designed that feature?
  • Based on how Gigabyte is handling this they should changes their name to "GigabyteMyShinyMetalAss".
  • It used to be worse. At an IT dept I worked for years ago, we had multiple brands and multiple supplies (not to mention motherboards) that would explode. One model would fail so bad it sounded like a gunshot.

Things are not as simple as they seems at first. - Edward Thorp

Working...