MSI Motherboard caught fire! What really happen?

I

ins-

[H]ard|Gawd
Joined
May 24, 2002
Messages
1,153
I recently got my computer built for a week now. 1200ax psu / msi 890fxa motherboard / h70 / 1090t so forth. (check sig for more)

Just today, I had my computer PRIME tested, to make sure a simple 4.0ghz overclock was stable. 30 minutes into the test, my computer shuts off. (computer ran for 1 week from now) I tried to turn it on. It went on then went off again. Then removed both side panels to see what was the deal. Nothing can be found. Then I smell something burning, then POOF! a flame errupted right close to the 12v motherboard line, and close to the northbridge. Luckily I reacted fast enough and unplugged the power and blew off the flame.

Here are some quick pictures.
2q8wva9.jpg

34qqqso.jpg

w2kndf.jpg


And no, the wires in the back was not touching the board, I have them nicely zipped tied. I just ripped it out, after the fire.

Let me know what you think. I have a new motherboard and cpu on the way. I just want to make sure its not the PSU or anything else, not sure how common this is, but it surely freaked me out. As long as I've been around computers, its a first. Let me know what you guys think.

I still don't really know what happen? Is is the PSU / MOBO? / or something else?
 
WOW. Looks to me like the VRMs overheated. Aftermarket VRM cooling possibly?
 
Bad MSI design as they're the most common, but the general consensus is that an OC'd Phenom II X6 pulls a hell of alot of power, putting strain on PWM/MOSFETS that were only designed to handle an OC Phenom II X4.
 
MassiveOverkill said:
Bad MSI design as they're the most common, but the general consensus is that an OC'd Phenom II X6 pulls a hell of alot of power, putting strain on PWM/MOSFETS that were only designed to handle an OC Phenom II X4.
Click to expand...

Eye-C says the blind Old Hippie.

Looks like a little extra cooling might be needed for those heavy overclockers. :)

It would have been nice if they had caught this flaw before production. :(
 
Cooling won't help enough it isn't designed to supply enough power to support the overclocks he was reaching. Dunno why no on has put out a 200w board on the market yet.

Hopefully you didn't damage the 1090t in the process.

Were you letting the cpu underclock or even run stock settings while running at idle?
 
Cooling won't help enough it isn't designed to supply enough power to support the overclocks he was reaching.
Click to expand...
Bummer!

Looks like there's no fixing this problem except going back to the drawing board?
 
holy shit! Makes me think twice now about leaving my computer running while stress testing. Glad you caught that and didn't burn down your house.
 
Old Hippie said:
Bummer!

Looks like there's no fixing this problem except going back to the drawing board?
Click to expand...

Well if you buy a different manufacturer's board, get the highest number of power phases possible. While the design might not be rated higher, the increased number of phases adds redundancy and robustness to the design and should make something like this much less common.
 
R-Type said:
Well if you buy a different manufacturer's board, get the highest number of power phases possible. While the design might not be rated higher, the increased number of phases adds redundancy and robustness to the design and should make something like this much less common.
Click to expand...
Totally agree.

This is the type of thing that happens when you're building to a price point and trying to incorporate many features without going over the price point.

Still, somewhere down the line somebody had to think/say "that's a lotta juice for these border-line components/configuration and it may cause a problem".

If it could be shown they (MSI) knew of the flaw and somebody's house burnt down because of it I'm sure there could be a legal consequence.
 
well, this isn't a good sign... I just bought an MSI 785GTM-E45 and had plans to upgrade it to a Phenom X6 in the future.
 
RogueTrip said:
Cooling won't help enough it isn't designed to supply enough power to support the overclocks he was reaching. Dunno why no on has put out a 200w board on the market yet.

Hopefully you didn't damage the 1090t in the process.

Were you letting the cpu underclock or even run stock settings while running at idle?
Click to expand...

Turbo Core was turned off. All core was running at 1.4 volt at 4ghz. Was stress testing it to see how well it will hold up. Long and behold, after 20-30 minutes of testing, this happens.

Thinking about forming a HOLY HOT MOTHERBOARD FIRE CLUB, anyone?

I had newegg overnight me a new motherboard and processor. I didn't want to hassle to find out if the processor was ok or not. Of course, free of charge on their end. I will test this once more, if it burns down, I will ask newegg for a refund, regardless of their policy. I'm sure I can demand it.
 
ryan_975 said:
well, this isn't a good sign... I just bought an MSI 785GTM-E45 and had plans to upgrade it to a Phenom X6 in the future.
Click to expand...

I've killed two of those in the last few months. Once with an x6, once with an x4
Posted via [H] Mobile Device
 
jebo_4jc said:
I've killed two of those in the last few months. Once with an x6, once with an x4
Posted via [H] Mobile Device
Click to expand...

were you overclocking? All I'm going to be running in it right now is an X2-5000+ and since it's my kid's computer, I don't think there's going to be any overclocking... now if they died running stock... shit.
 
ryan_975 said:
were you overclocking? All I'm going to be running in it right now is an X2-5000+ and since it's my kid's computer, I don't think there's going to be any overclocking... now if they died running stock... shit.
Click to expand...
Ironically, the X6 was overclocked, the X4 was running stock.

ins- said:
Did you have the board replace? How is it running as of today if so?
Click to expand...

Long story short, I started with the 785-e45 and an unlockable X2. It ran fine with 3 cores enabled, overclocked around 3.8ghz, for a few weeks. Then I swapped out the X2 for a 1055t. It ran the 1055 fine at 3.6ghz for a few days, then I pushed it up to 3.8 and it blew a cap within minutes. I figured the MSI just wasn't built to handle a 3.8ghz Thuban, so I swapped it out for a Biostar 890GXE and it's working great to this day.

The second time I bought the E45 with an Athlon II X4 630 combo at Microcenter. It blew some sort of component while stress testing at stock speeds. I replaced it with an ASUS 880G mobo.

So, the first board was probably fine for an X2 or X4 but it was stressed too hard. The second board was a POS any way you slice it. Microcenter is the real winner here. I always start with the low end mobo then end up upgrading to a higher end one...
 
ins- said:
Turbo Core was turned off. All core was running at 1.4 volt at 4ghz. Was stress testing it to see how well it will hold up. Long and behold, after 20-30 minutes of testing, this happens.

Thinking about forming a HOLY HOT MOTHERBOARD FIRE CLUB, anyone?

I had newegg overnight me a new motherboard and processor. I didn't want to hassle to find out if the processor was ok or not. Of course, free of charge on their end. I will test this once more, if it burns down, I will ask newegg for a refund, regardless of their policy. I'm sure I can demand it.
Click to expand...

voltage sounds low for a 4ghz OC....


onto the problem----MSI has had a bad track record since the hexes came out, atleast when it comes to OC'ing the x6's......

For example, i've got a Crosshair III, no issues at all, 790FX, its well built, some of the MSI's are not like this, they are shoddily built(again in comparison to the Crosshair III a higher level enthusiast board)....


I think this is the first MSI 890FX(their enthusiant one i believe) that has blown thouhg.....probably a bad board in this case, htough i dont know if I would trust one after what you have gone through
 
I've had a my Thuban running in my at the settings listed below in my 790FX-GD70 for about 3 months now and I've not had any issue's. I am running the same CPU in a 890GX-G65 @ 3.8ghz 1.4v and it's been chugging along fine stuffed in a SG03 with the stock HSF and a 5870.
 
wow,glad you caught and didnt end up with a house fire or something
 
Awesome, I guess its a luck of draw.

Yea, I've been pushing the voltage slowly at 1.4v and it "was" stable for 20-30 minutes on Prime, until it shutted off. LOL. Everything else was leave stock, besides the multiplier at 19? 20?

But yea, I thought the 890fxa-gd70 was the enthusiest line as well, where things like this won't happen. Yet it does... Hmm, Yes, Im a little wary on what I should do from this point. I will try and OC the same thing on the new board/cpu. If it does this again, I will resort to another manufacture board.
 
I've been shopping AM3 boards recently and I've seen a number of reports of MSI boards catching on fire. I didn't pay a lot of attention to the details but at least one mentioned vrm's. I don't think all of the cases involved X6's.
 
I'm 58yrs old and don't consider our legal system a joke nor do I condone frivolous litigation but this looks like a pretty blatent oversite considering the possible outcome from this design flaw.

I'm sure MSI would point to the fact that overclocking isn't encouraged and the boards only reacted to the abuse but....it's still kinda scary.
 
Yep that's a 4+1 power phase design... it will do that...

that massive heatsink they put on the VRM, is quite a joke... you just can't cool a 4+1 power phase design when you are OC'ing an x6
 
Old Hippie said:
I'm sure MSI would point to the fact that overclocking isn't encouraged and the boards only reacted to the abuse but....it's still kinda scary.
Click to expand...

My MSI 870A Fuzion board has a Overclock Genie button (OC Genie) so it would be presumed the intent of this motherboard was specifically for overclocking.
 
EricFX1984 said:
Yep that's a 4+1 power phase design... it will do that...

that massive heatsink they put on the VRM, is quite a joke... you just can't cool a 4+1 power phase design when you are OC'ing an x6
Click to expand...

Huh.. my Gigabyte 890GPA uses 4+1 and has been running my x6 at 4Ghz for months now. Kind of scary to think that it could burn my house down. I have not read about this happening with any of the Gigabyte boards though.
 
Running at 4ghz isn't a problem. Try running it at 4ghz at all 6 cores, and stress test it at 100% load for say an hr? Hmm, no kidding, the outcome of this is almost a lawsuit waiting to happen. I wonder what MSI has to say about this?

And yes, this board has Easy Overclock Features built in. So, I beleive its suppose to do what its meant to do.
 
ouch 4+1 vrm? I am skeptical oc'ing quads with that setup. From a quick look on newegg, I know the Asus Crossfire III has 8+2 which should eliminate any of these sort of problems.
 
nailzer said:
My MSI 870A Fuzion board has a Overclock Genie button (OC Genie) so it would be presumed the intent of this motherboard was specifically for overclocking.
Click to expand...

No there's something that'd be hard to deny!
 
Does anyone plot the temperatures of the MOSFETs and coils vs. CPU clock, to see if the temperature rise starts to outpace the increase in clock speed?
 
I wonder should I just return the new motherboard and get the Asus ?
 
EricFX1984 said:
Yep that's a 4+1 power phase design... it will do that...

that massive heatsink they put on the VRM, is quite a joke... you just can't cool a 4+1 power phase design when you are OC'ing an x6
Click to expand...

The MSI 890fxa-gd70 seems to have a 4x4 power phase to me? Or am I mis-understanding something.
 
ins- said:
The MSI 890fxa-gd70 seems to have a 4x4 power phase to me? Or am I mis-understanding something.
Click to expand...

I wouldn't trust to that. I fried an MSI NF980-G65 with a 1090T, and as near as I can tell your board and this one have the same power phase. The MSI website, and even the manuals you can download for specific boards, are also woefully lacking this information (as well as a frustrating amount of other details).

Bottom line seems to be to steer clear of OCing a Thuban on an MSI mobo.

EDIT: here's the first post I made on the subject, in the thread previously linked.
 
Thanks for the post and warning. I was torn between an MSI board and ASUS board with almost the exact same specs. the MSI was cheaper but I will pay a little extra for my house not to burn down.
 
MassiveOverkill said:
Bad MSI design as they're the most common, but the general consensus is that an OC'd Phenom II X6 pulls a hell of alot of power, putting strain on PWM/MOSFETS that were only designed to handle an OC Phenom II X4.
Click to expand...

This pretty much sums it up.

rocks1982 said:
WOW... Ive never seen something like this.
Thank GOD.
Click to expand...

I have. The first Athlons did this sort of thing constantly. We had stacks of bad boards like this in the tech shop I worked at during that time.
 
This is a known issue with the 790fxa and 890fxa boards. I speak from personal experience.

The ultimate cause...the VRM heatsink doesn't sit properly flat and contact all VRM. The result, for you, is more severe than me. I just had a blistered board. MSI is NOT admitting any defects, and treats the problem like any other RMA. I swore off MSI because of this issue, and their handling of it.

If they fix the heatsink problem, then the problem is solved for overheating and catching fire.

The person above me isn't quite correct. MSI's FXA boards are advertised as supporting 140 watt processors. The 1090T is a 125 watt model. I had a C2 965 in my MSI for about 6 months (having to RMA it at the 3 month mark). It was a flaked BIOS chip. MSI replaced it, and I had the same board back. So for that 6 months I had the 965, which was overclocked to 3.8 Ghz (stock voltage). I'm noticing during about a week before I dropped the 1090T in that voltages were fluctuating a bit more than usual (this board is kinda bad for that). Days after popping in the 1090T the system crashes and PSU goes into protect. I had not been pushing stuff with extra voltage.

Once I took off the heatsink, I saw that one VRM hadn't even contacted it, which is the one that blistered the back of the motherboard.
 
Last edited:
You must log in or register to reply here.
Back
Top