Tuesday, December 27th 2011
Gigabyte Recalling X79 UD3, UD5, G1.Assassin 2 Motherboards
Last week, a Taiwanese overclocker putting his OC workbench through an relatively laxed OC stress test saw its Gigabyte X79 UD3 motherboard go bust. Its CPU VRM couldn't cope with the stress, and blew a MOSFET. At the time, people responding to his video condoled him for his bad luck. It appears now that his wasn't a one-off case of "bad-egg". Gigabyte, in its latest press release on its Chinese website, noted the issue. Apparently it received several such complaints from overclockers where even moderate voltage-assisted CPU OC fried its VRM. The issue was found to be widespread, among three of its main socket LGA2011 products, the GA-X79-UD3, GA-X79-UD5, and G1.Assassin 2.
Apparently, the issue is caused by a cocktail of bad firmware to complement the board's PWM circuitry, and bad quality PWM components. As an immediate remedy, Gigabyte issued a BIOS update for the affected products. This BIOS, however, will cripple the board's overclocking abilities. The new BIOS will throttle CPU when subjected to extreme stress, to save the VRM. The BIOS remedy is only for those who opt to keep their boards, or don't subject the board to extreme tuning. The other remedy, is to return the board to Gigabyte, for a free replacement when the "right" boards are available. Gigabyte also announced a general recall of the GA-X79-UD3, GA-X79-UD5, and G1.Assassin 2, from the market. A video of the "unlucky" (not anymore) overclocker's day going bad, can be watched here.
Update (29/12): Gigabyte's German office wrote to us and explained that on their end they find the problem to be because of bad firmware, and not bad component quality; and that unlike Gigabyte Taiwan, they are not recalling products or soliciting replacements, but asking users to update their BIOS to the latest available. Gigabyte Germany set up a hotline for German customers, that's 040-253304-55.
Source:
Gigabyte
Apparently, the issue is caused by a cocktail of bad firmware to complement the board's PWM circuitry, and bad quality PWM components. As an immediate remedy, Gigabyte issued a BIOS update for the affected products. This BIOS, however, will cripple the board's overclocking abilities. The new BIOS will throttle CPU when subjected to extreme stress, to save the VRM. The BIOS remedy is only for those who opt to keep their boards, or don't subject the board to extreme tuning. The other remedy, is to return the board to Gigabyte, for a free replacement when the "right" boards are available. Gigabyte also announced a general recall of the GA-X79-UD3, GA-X79-UD5, and G1.Assassin 2, from the market. A video of the "unlucky" (not anymore) overclocker's day going bad, can be watched here.
70 Comments on Gigabyte Recalling X79 UD3, UD5, G1.Assassin 2 Motherboards
It's just that at the moment they're soliciting free replacements from "Taiwan, Penghu, Kinmen and Matsu area". Soon they'll extend it to every other territory they sell in. They'll have to.
Yes, your GA-X79-UD5 is plotting to kill your CPU the same way its siblings from "Taiwan, Penghu, Kinmen and Matsu area" are.
Well I was wondering if perhaps they identified the problem to a specific batch of boards that were distributed in that area, or something.
I mean, if that's the case, I need to contact Gigabyte and get a replacement myself. :laugh: I pushed 265W through my board's VRM for 5 hours, and my board is fine. So I must assume it's near death now then. I saw that video of the UD3 burning up last early week.
And yes dave, let the UD5 die muhahaha, i already got rid of mine.
Never freaking settle (If an asrock employee read this, i want some market share k10xbi)
Commonly 56x+ CPU's will bring far more than retail but 10x is not remotely common or paid for. 3-5x sure.
As far as who is responsible.. is giga supposed to test to 6Ghz? 5Ghz? 4.5Ghz? 150W/250W/350W? I hold giga responsible for this b/c of the shoddy performance and parts, but the testing part... I mean where (power level) does it happen? Is this part of their testing procedures to test that high?
EDIT: I did have a UD7 die mid review. It was benched hard (4.8Ghz, with peak 4.949Ghz), but then was set to stock speeds, left on for 2 days, then in the was dead.
Don't forget them faking PCIe lane counts on G31 boards, overheating 780G northbridges due to crappy heatsinks, or using tricks to make their 1155 boards PCIe 3.0 capable with PCIe 2.0 switches on the boards. I'd rather use MSi over them if I had to (seems like they've improved their standards).
I'll stick with ASRock/ASUS for now. My ASRock Z68 Extreme4 Gen3 is solid, my ASRock A770DE+ is a $53, practically unkillable board (runs 1.5v/3.9GHz 955BE + 5770 + 6670 non-stop folding/crunching), and my old Crosshair III Formula is retired currently after 2.5yrs, but is still as solid as it was from day one. I'm a bit surprised my dad's Gigabyte 790GX board and friend's Gigabyte P55A boards still work (although plugging a cellphone into the front USB headers on my friend's P55 board causes the computer to immediately shut off).
The issue is a compound one, I'm not blaming Intel or using that as an excuse, I'm saying it didn't help the situation.
The real blame lands solely on the manufacturer of the mosfets that didn't meet spec.
The rumors about PWM issues where there for a long time, especially with with the boards missing the X79 launch, having no stock available for a launch like this isn't a small hick up.In the video I saw fire.. that would be more accurate :p
But like other guys said, this is clearly a case of rushing products to market and lack of sufficient testing. I know they do that with their video cards, cause coil noise runs rampant on their higher end products... Plus an entire series of early 6850's that were all duds... Black screen, wrong fan setup in the default BIOS. Too bad it's catching on to motherboards now.
And yes, pulling everything when the system crashes was very ninja, certainly not my first thought.
My theory is, ASUS paid Chil to sabotage IR taking CHil's firmware.
FYI is GIGABYTE didn't believe in their products they would have never offered 5 year warranties on the UD3, UD5, G1 Assassin 2, and UD7 in the USA.
If they have a part that is just marginally out of spec a good amount of samples will pass, likely enough to pass quality control, and a fair amount will go up in smoke when pushed hard.
If you check Gigabyte site, earlier BIOSes are no longer listed, and the affected products all have Bios ver. "F7" listed.
Just my thoughts. ;)
I doubt it just needs a bios flash, its a major recall.
1: Boards with specific model BIOS are recalled from retail shops.
2: Users that already bought the board can flash to the "F7" BIOS, and boards now have lifetime warranty(instead of 5 years in NA)
3: Users that are still not comfortable with the BIOS flash have unconditional return policy.
I've touched base with Gigabyte to get clarification on the issue. I'll update as info comes in, but in the meantime, it seems that Gigabyte has found fault in the BIOS itself, and not the hardware. I doubt they'd extend warranty to lifetime on products that have faulty hardware. Remains to be seen if the BIOS update affect clocknig or anything, but as I am also working on a review of the GA-X79-UD5, you can bet that I'll check it out.
So basically you can flash the crippled f7 bios and live with it, which makes overclocking worse or return it if you're not happy?