Question Wow...I'm stunned...Gigabyte power supplies exploding. Sounds like one major vendor is trying to unload them in bundles.

[BoomerD]

I don't know who makes these PSU's for Gigabyte, but I'd think Gigabyte would be on the hook for damages caused by these "exploding" power supplies.

 

[UsandThem]

Gigabyte has historically had some of the worst power supplies out there , where I'd never consider buying at any price.

For some reason, they have been OK with their crap power supplies not doing their brand name any good. You'd think they would want to only associate their brand with quality products, but nope.

They must use the cheapest OEMs they can find, and then simply slap their name on it.

 

[UsandThem]

Someone on Anandtech's main site recently linked to a "deal" for a Gigabyte power supply, I guess only using the $30 price drop as their gauge for being a good deal: https://www.anandtech.com/show/16845/at-deals-gigabyte-850w-psu-gpp850gm-lowers-to-109

Thankfully someone in the comment section linked to a recent review where the power supply, yes you guessed it, literally blew up while being tested: https://www.techpowerup.com/review/gigabyte-gp-p750gm-750-w/

Despite that being bad enough itself, the reviewer then stated what Gigabyte's "response" was to it, and that was even more shocking to me. You can tell that this must be at least a semi-frequent occurrence because they didn't even want the unit to be sent back to be evaluated.

Of course, I sent Gigabyte an email with all of my findings and spoke to the power R&D supervisor, who informed me that they checked five samples which all survived their OPP evaluation test. What left a negative impression on me was that they didn't bother asking for my sample to be sent back to check on what went wrong. With such a colossal failure on hand, the respective brand usually immediately asks for the sample to be returned for a closer examination to find the source of the problem. You might test dozens of samples at the factory, and while all of those pass the tests, what matters the most is finding those with an actual problem for a thorough examination in an attempt to figure out why they broke down.

 

[Leeea]

I don't know who makes these PSU's for Gigabyte, but I'd think Gigabyte would be on the hook for damages caused by these "exploding" power supplies.

It is worse then that.

If the leadership at Gigabyte is Ok with doing this to their power supply customers, they are ok with doing this to all their customers.

I would not buy anything with the Gigabyte brand on it.

 

[VirtualLarry]

They make not-so-great RTX 3090 cards, so I hear. If used for mining, the Gigabyte cards are manditory to replace the (poor-quality) factory thermal pads, so that your card doesn't burn up the VRAM.

 

[BoomerD]

I've never heard anything good about any Gigabyte product except their motherboards...and that's why I've never bought one of those either. If poor to mediocre quality is OK with the company for the other categories...why not in their mobo lineup?

 

[UsandThem]

I've never heard anything good about any Gigabyte product except their motherboards...and that's why I've never bought one of those either. If poor to mediocre quality is OK with the company for the other categories...why not in their mobo lineup?

Gigabyte was pretty much my go to for motherboards ever since I began building my own back in the 90s (as well as Asus).

That changed around 2015 when I built my son's previous computer. I ordered a pretty mid-range Gigabyte H97 motherboard. Once it arrived, I preceded to spend days swapping components in and out because of issues. I eventually replaced everything with pulled components from other computers, and it still was crashing.

I sent the motherboard back to Amazon, and they sent a replacement. It was the exact same problem. I then Googled the Gigabyte model number, and found there was an abhorrently high number of people on various forums who had serious issues with that particular motherboard, and it really seemed like that particular model had a serious design flaw, and it really should have been recalled by Gigabyte. It was a particular model that a lot of people initially used to build "Hackintosh" computers over on the TonyMac86x forums, but many eventually gave up and replaced it with a different model.

I then ordered an Asus motherboard instead, and returned the 2nd Gigabyte. It worked flawlessly the 1st time. All companies have a dud every now and then, but to me it seems like Gigabyte simply looks the other way and continues to sell components that really should be recalled.

 

[BoomerD]

Gigabyte was pretty much my go to for motherboards ever since I began building my own back in the 90s (as well as Asus).

That changed around 2015 when I built my son's previous computer. I ordered a pretty mid-range Gigabyte H97 motherboard. Once it arrived, I preceded to spend days swapping components in and out because of issues. I eventually replaced everything with pulled components from other computers, and it still was crashing.

I sent the motherboard back to Amazon, and they sent a replacement. It was the exact same problem. I then Googled the Gigabyte model number, and found there was an abhorrently high number of people on various forums who had serious issues with that particular motherboard, and it really seemed like that particular model had a serious design flaw, and it really should have been recalled by Gigabyte. It was a particular model that a lot of people initially used to build "Hackintosh" computers over on the TonyMac86x forums, but many eventually gave up and replaced it with a different model.

I then ordered an Asus motherboard instead, and returned the 2nd Gigabyte. It worked flawlessly the 1st time. All companies have a dud every now and then, but to me it seems like Gigabyte simply looks the other way and continues to sell components that really should be recalled.

ASUS motherboards have never failed me. I do tend to get their better (Not top of the line RoG) boards rather than their cheapest models

 

[Steltek]

Gigabyte used to source a lot of their branded PSUs from CWT, Andyson, FSP, and Seventeam. They weren't selling the best quality PSUs, but they also weren't RAIDMAX levels of junk either.

However, I think a lot of their current ones are made by an Chinese OEM named MEIC, which apparently had never even produced PSUs before (they used to mainly make OEM power adapters of all things). The MEIC company website is gone now -- I guess it turns out selling products that blow up right and left all over the place is not conducive to good PR....

 

[UsandThem]

It looks like Gigabyte is having a REALLY bad week. First, news of their PSUs exploding are spreading across hardware sites, and now their headquarters' server was successfully attacked:

https://www.tomshardware.com/news/gigabyte-ransomware-hack

According to the report, attackers stole as much as 112GB of confidential data, which includes motherboard designs, sensitive encryption keys, UEFI BIOS versions for yet-unreleased products, TPM data, and much more. Gigabyte is now in danger of all of this information making its way to the public domain. To keep that from happening, the hackers are requesting a ransom, as they're threatening to leak the data to the public and compromise Gigabyte.

On a positive note, it appears that Anandtech's main site scrubbed all existence of the Gigabyte PSU "deal" they posted (dead link in post #3).

 

[UsandThem]

When you dig in your heels, and decide to make the problem even more embarrassing for their brand reputation:

https://www.tomshardware.com/news/gigabyte-releases-statement-on-exploding-psus

Mpitziopoulos concluded that the OPP and OCP at 12V protections were set too high and the questionable FETs simply couldn't deal with the load. Even with his years of experience in the field, the GP-P750GM utilized parts from brands that Mpitziopoulos had never heard about, which make us doubt the quality of the components.

Gigabyte's solution consists in lowering the OPP trigger point for the GP-P850GM and GP-P750GM. The manufactuer reduced the threshold from 120% to 150% down to 110% and 120% for both units. In its statement, Gigabyte reaffirms that "the potential issues that were reported, only seemed to occur after very long time periods of extreme load testing via DC Electronic load equipment and would not be typical of any real world usage."

 

[DigDog]

Gb has lost whatever brand reputation they had to the gray hand of time. I can't think of any Gb product after the Conroe years that stood out.

 

[bononos]

They make not-so-great RTX 3090 cards, so I hear. If used for mining, the Gigabyte cards are manditory to replace the (poor-quality) factory thermal pads, so that your card doesn't burn up the VRAM.

Sounds like the brand is just coasting off their reputation at this point. Such a shame really.

GN bought 10 of the problem psu's and tested them -

 

[DeathReborn]

I have built 3 systems for clients using these PSU's, suffice to say I have offered them a replacement with Seagate, EVGA or Corsair units. Going to replace one tomorrow, the others haven't responded yet & I won't be buying any more GigaBoom products for a good while.

 

[mikeymikec]

only seemed to occur after very long time periods of extreme load testing via DC Electronic load equipment and would not be typical of any real world usage."

Reminds me of Microsoft's "that vulnerability is purely theoretical!" quote.

 

[Magic Carpet]

They make not-so-great RTX 3090 cards, so I hear. If used for mining, the Gigabyte cards are mandatory to replace the (poor-quality) factory thermal pads, so that your card doesn't burn up the VRAM.

This problem is not exclusive to the Gigabyte brand. When used for mining 24/7, most 3090s need some kind of tweaks. Out of the box, EVGA/ASUS/PALIT seem to be the best in this department, though. It sucks to see your hashrate going down because of memory throttling. 3060 Ti's are better for mining, imo

3090 is best for content creation and/or gaming, but not mining.

These tweaks I am talking about

 

[UsandThem]

https://www.tomshardware.com/news/p...esponds-to-gigabyte-exploding-psu-shenanigans

Aris responds to Gigabyte's latest actions/claims.

Instead of fixing the issue and putting out the fire they caused by using crap components and/or a bad design, Gigabyte is really taking a very bad PR hit over this mess.

Going forward, it will be extremely unlikely I would trust in any of their products. I hope when their ownership looks back on this period, they conclude their tone-deaf stance was the correct choice.

 

[coercitiv]

Excellent PR strategy to draw attention away from their ransomware problems.

then again, angry Steve is entertaining... so thanks GB!

 

[mikeymikec]

I think that second video should have been more strongly worded with respect to what I'd describe as a recall for literally no reason. Gigabyte are claiming that the testing is faulty, not the product, but they're going to do a recall anyway. That's how ridiculous their position is.

 

[witchapon2530]

A lot of time people put it on the OEM, whereas every OEM can pretty much doing as good as they were paid for.

I watch tomhardware PSU reviewer Dr. a... something, he tear down the PSU and find it very good quality build (from this OEM) and he had the same issue as Steve and suspect the cause to be in Gigabyte design, so in this case it's on Gigabyte I guess.

 

[DAPUNISHER]

The PSU having evidently failed to light a big enough fire under Gigabyte's butt, on its own. The bad press fanning the flames did the trick. They fixed the issue, in the latest version -

 

[DeathReborn]

The PSU having evidently failed to light a big enough fire under Gigabyte's butt, on its own. The bad press fanning the flames did the trick. They fixed the issue, in the latest version -

To me that name is toxic, (P750GM etc, not GigaByte) so I won't touch them ever again.

 

[UsandThem]

Aris received a fixed unit in to review, and despite Gigabyte's claims that the reason was because the OPP and OCP protections were set too high, it appears they had the OEM do some changes to the PCB:

https://www.tomshardware.com/reviews/gigabyte-gp-p750gm-power-supply-review

On the left is the old platform and on the right is the new one. The difference in design is located near the LLC resonant controller. The resistance of R94 has decreased, and the layout is also different. R94 is connected to the ILIM pin of the CM6901 controller, which is responsible for OPP's triggering point. If MEIC just wanted to lower OPP, there was no reason to change the layout in this area. A simple resistor replacement would do. Apparently, the company needed to tweak the design, so it decided to change the PCB.

However, this still should be a unit most "in the know" shoppers avoid as there is no way to know if they'll get the fixed (or the exploding) version.

Not to mention, even with the updates, it's still a loud and poorly performing unit anyways.