GIGABYTE's mobo X79-UD3 happens self-ignition and burning!

[Eldonko]

G.Skill has told me that they dont have Trident anymore so if there is new stuff I demand to know where to get it and the specs!

[sin0822]

Did you hear that he got the board replaced with and Assassin 2 and new CPU(he went from 3930K to 3960X!) and memory. Maybe i am a bit behind.

[zsamz_]

Did you hear that he got the board replaced with and Assassin 2 and new CPU and memory. Maybe i am a bit behind.

so if i get a ud3 n the pwm blows up i get an Assassin 2 ?

gettin the heatgun n cameras ready lol

[Hondacity]

so if i get a ud3 n the pwm blows up i get an Assassin 2 ?

gettin the heatgun n cameras ready lol

you're in canaduh, gigabutt will take forever to respond to your email lol

if you were in taiwan, gigabt HQ will send 3 engineers with gifts lol

[BeepBeep2]

I don't know anything about International Rectifier, but I do know that Gigabyte, Asus and MSI have started using cheap "NIKOS" transistors even in their high-end motherboards. You can't trust these people, they're very deceptive, they have to guarantee SOME failure out there in the field to ensure further sales.

My point is that you don't see these types of failures with ST Microelectronics and such transistors.

Yea, IR is probably an OK transistor manufacturer, probably superior to the cheap NIKOS garbage they all use in their lower-end boards, but why do I always get the feeling that the Chinese and Taiwanese are always penny pinching, even on expensive, high end products. And then they complain about desktop sales. They have destroyed the desktop market with their inferior products. First they were nickel and diming us on the capacitors, and now this funny business with transistors.

How much extra would it cost for STM transistors? A few pennies? Save the cheap stuff for the cheap boards. We don't need tantalum capacitors in the VRM MSI (Big Bang series), what we need are decent transistors.

Do you know what a transistor is?...here, we are talking about a Metal Oxide Field-Effect transistor, or (MOS)FET.
Please, don't talk about things you don't know about. I'd know, I'm extremely guilty of that as well.

You know CHiL? The voltage controller used on all of AMD/NVIDIA's reference boards? Yeah, they got bought out by IR a long time ago. Gigabyte has used IR for a long, long time on their boards.

Have you tested all the top models of each manufacturer's boards to find these "NIKOS" Power MOSFETs? No?...didn't think so.

The problem here is not the MOSFET. It is the OCP and OTP circuits to blame. The board should throttle the CPU or even simply shut down when the VRM section gets too hot. This board didn't shut down until it caught fire.

[STEvil]

Do you know what a transistor is?...here, we are talking about a Metal Oxide Field-Effect transistor, or (MOS)FET.
Please, don't talk about things you don't know about. I'd know, I'm extremely guilty of that as well.

You know CHiL? The voltage controller used on all of AMD/NVIDIA's reference boards? Yeah, they got bought out by IR a long time ago. Gigabyte has used IR for a long, long time on their boards.

Have you tested all the top models of each manufacturer's boards to find these "NIKOS" Power MOSFETs? No?...didn't think so.

The problem here is not the MOSFET. It is the OCP and OTP circuits to blame. The board should throttle the CPU or even simply shut down when the VRM section gets too hot. This board didn't shut down until it caught fire.

Actually in this instance it was the fet that was to blame. It was a faulty fet, they happen.

It would be nice to have a VRM measuring system that actually worked though. None that I am aware of monitor each fet/driver IC and only measure PCB or heatsink temperatures at the most with reference to temperature. Bad fets often enter thermal cascade long before the PCB or heatsink temperatures reach the thermal tripping point.

[zsamz_]

you're in canaduh, gigabutt will take forever to respond to your email lol

if you were in taiwan, gigabt HQ will send 3 engineers with gifts lol

your right
it took them almost 3 months to rma my p67 b2 board

was like pullin teeth to get a p67 rma from gigabut only got the runaround

bought a gene-z
sold the ud7 the next day for cheap

wish i made a video
they woulda sent stripers a g1 assasin and a 6990 lol

i rather have a gene-z

my freind Radaja won a prize a gigabut board

over 3 months he has yet to get the board or other goodies n he in the US lol

when i was workin in china i use to get freebies daily lol

[BababooeyHTJ]

hmm...

I see nobody has commented about how USELESS the VRM heatsinks usually are.

They often make poor contact and the TIM can act as an insulator if its too thick or the motherboard bows due to poor isntallation methods.


With that said, the fault lies likely with a faulty component as there is not a rash of them blowing up yet, unlike the Asus Z7S-WS for example of which there were many that did blow up due to the poor VRM design.

I completely agree with you. I really loved my P5Q Deluxe with backplates on the vrm sinks. Thats something that I wish was done more often. Its too bad that stock aluminum heatsinks still sucked. I don't understand why these little aluminum sinks. Once I bolted on a couple of enzotech mst-88 sinks to the stock backplate that was the perfect solution.

[STEvil]

best is individual heatsinks, or fets bolted to heatsink (used to see that back in the day..) but its less efficient on the manufacturer end, so we put up with crap :/

[Particle]

I guess I have three things to say, after having read the thread to this point:

1) The guy wasn't being smart. He failed to cool the VRM when using water cooling. If a board he uses like that catches on fire, he should be punished by being denied replacement instead of receiving free upgrades. This guy isn't going to learn that what he did was stupid.

2) In this PARTICULAR case, I have to agree that it's possibly a bad component.

3) For those of you defending him because temperature protection circuits are generally present, you're not standing on good ground. It's not good practice to count on an emergency system to be routinely used in general operation. That alone should make you re-examine your perspective. Since they're popular according to my sig, let's throw in a car analogy! Counting on VRM high-temp shutdown is like driving without breaks because hey, your airbag will save you, right? Just like when it comes to the electronics this analogy is tied to, the answer is "Yeah, maybe." One thing it definitely isn't going to save you from is an excessive peak transient. A FET will just blow up and the protection circuit would have insufficient time to react.

[Kallenator]

A motherboard which supports CPU's where one of the standard cooling solutions is water NEEDS to have a viable passive cooling solution on their MOSFETs. And I am sure they have, this component as sin0822 so kindly linked the datasheet to, can take anything from -55 - 155 degrees Celsius and operate normally. It does not however have a built in fail safe for over heating. (As far as I can see at least) However you will find this function in more advanced MOSFETs and VRMs like the 78xx series.
So the motherboard needs to have some surveillance on this, and it does, you can check VRM temperature in the BIOS, however it does not have a temperature sensor on each MOSFET so if one gets out of line it can't necessarily throttle back.
Gigabytes fault? Hardly... components fail, it happens, its rare, but this just gets way too much attention.

Datasheet:
http://www.irf.com/product-info/data...rfh8330pbf.pdf


Oh, and Merry Christmas everyone! =D

[Warwian]

At least it looks like Gigabyte is taking this issue very seriously as they've started recalling some of their X79-boards. And as a temporary solution to these problems they've released an update to the UEFI/BIOS which limits overclocking and throttles the CPU during stress.

http://www.gigabyte.tw/press-center/....aspx?nid=1077

(translated: http://translate.google.se/translate...x%3Fnid%3D1077)

Swedish online magazine Sweclockers just posted about it here:

http://www.sweclockers.com/nyhet/148...och-assassin-2

(translated: http://translate.google.se/translate...och-assassin-2)

Edit:
It looks like lots of sites have already picked up on this:

http://vr-zone.com/articles/gigabyte...rds/14360.html
http://hwbot.org/forum/showthread.php?p=144646
http://www.techpowerup.com/157543/Gi...herboards.html

[Hondacity]

just affirms my statements, "stupid gigabutt product" and "recall" lol

[matari]

Gigabutt got caught using cheap components. Wonder who else is too.

[Hondacity]

why would a manufacturer not stress test their boards? its just stupid. selling these in thousands and not testing it(bios and hardware)

[zsamz_]

just affirms my statements, "stupid gigabutt product" and "recall" lol

+1

@ least this time they recalled a product

in the past people stuk in rma loop

like my rma loop with p35c-ds3r
6 rmas in 8 months

[mockingbird]

What are they going to replace those IR HexFETS with?

[HelixPC]

Gigabyte Recalling X79 UD3, UD5, G1.Assassin 2 Motherboards

[Origin_Unknown]

hold on, the board they gave him to replace the faulty one is also in the recall list? lol

[ChinStrap]

Wow. Someone in their PR dept. is having a bad day.

[sin0822]

I only read BIOS recall, am i alone or what? I don't see board recall.

[sin0822]

What are they going to replace those IR HexFETS with?

They didn't say anything about the X79-UD7 did they? Then what makes you think they will recall the FETs. THis issue has nothing to do with the MOSFETs.

[defect9]

I only read BIOS recall, am i alone or what? I don't see board recall.

from the article: "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."

[sin0822]

I am talking about the official statement, website article i don't put any credit into, they got their info from the same place.

[Warwian]

They didn't say anything about the X79-UD7 did they? Then what makes you think they will recall the FETs. THis issue has nothing to do with the MOSFETs.

I don't have a great deal of insight into this story but it seems like there might be firmware issues with the PWM on the UD3, UD5 and the Assassin 2. If this is a quality issue or not is of course hard to say, but I remember reading something about there being suspicions of bad quality PWM components also playing a role in this. Anyway, I trust International Rectifier's products (which I believe they use on the UD3, UD5 and Assassin 2?), whether it's the new IR digital PWM or the IR MOSFETs. And I'm convinced Gigabyte will be able to continue using them and that this'll be sorted soon without much of a hassle.

By the way, I find it strange that the UD7 is seemingly unaffected by all this. Doesn't the UD7 use mostly the same digital PWM/components for the VRM as the other boards? Probably not as that would be expensive! :o