• Some users have recently had their accounts hijacked. It seems that the now defunct EVGA forums might have compromised your password there and seems many are using the same PW here. We would suggest you UPDATE YOUR PASSWORD and TURN ON 2FA for your account here to further secure it. None of the compromised accounts had 2FA turned on.
    Once you have enabled 2FA, your account will be updated soon to show a badge, letting other members know that you use 2FA to protect your account. This should be beneficial for everyone that uses FSFT.

7950's caught on fire!

S

Sarian

n00b
Joined
Jun 28, 2005
Messages
57
I have a machine setup with 4x Gigabyte 7950's that I've been litecoin mining with for a few months. It strangely crashed one day. I turned it back on. Sparks flew and a nice red glow was left over while I flailed at the power switch to cut the power. Here's the result.

http://i.imgur.com/8vQ4hi9.jpg

So I pulled the card, submitted a RMA and sent it in. Continued mining again with 3 GPU's. Fast forward another 2 weeks and we have another crash! I start it back up again and the same thing happens. Card #2.

http://i.imgur.com/S1P9d88.jpg

Now could this be a card defect or am I dealing with a PSU problem. I really don't want to keep lighting cards on fire. It's not like I skimped on the PSU (SeaSonic X-1250).

Does anyone have any insight?
 
This thread is smokin' hot. :D

Bad puns aside, I can vouch for Gigabyte being pretty good about honoring their RMAs. It's why I always try and buy their boards and WON'T touch ASUS with a 10 foot pole.
 
Fortunately they've approved my first RMA... I haven't done the second one yet.

I've developed this learned behavior of turning that computer on = bad. I'm scared of another one burning up.
 
I would contact them and ask them to verify the quality of your remaining cards given the same very bad issue has cropped up twice in a short space of time.

Before doing that, can you get a voltmeter on your PSU to check the 12V and 5V rails when idle and under load?

Mr. Bluntman said:
This thread is smokin' hot. :D

Bad puns aside, I can vouch for Gigabyte being pretty good about honoring their RMAs. It's why I always try and buy their boards and WON'T touch ASUS with a 10 foot pole.
Click to expand...
Same here, I had the same issue with the CPU socket on both Asus and Gigabyte boards, Asus refused warranty, Gigabyte replaced the board.
They've already won my next purchase.
 
I had the exact same issue with 2 msi 7950's.
It happened in almost the same location.

it looked like the fault was the underside of my cards at the 3rd capacitor from the bottom of the card.
This occurred when I did not have auto-gpu turned on. Since then I have turned auto-gpu on and haven't had any problems. I have 2 computers with 4x7950's each and 2 comps with 3x7950's each and I'm using PCP&C 950w psus in each comp. Yes, a 950w psu powering the 4x7950's. However, I have my cards undervolted to 1030 to 1050mV (reduced voltage a little more since this incident).

The 7950's that fried on me had dual 6-pin connectors. All my 8pin cards are fine.
I love MSI. They replaced the cards without giving me any problems. I didn't have to make a thread to turn some heads in order to get MSI to stand behind their warranty. It just took about 8-10 days before they shipped replacements.

Did you have auto-gpu turned on? What core/mem speeds? voltages? temps?
I've set temp cutoff at 93C now, target 84C. Using my IR temp laser while mining I found this area on the back of the PCB is the hottest part of the gpu.


J Macker said:
wow. That sucks. Sorry OP.
On a similar note, MSI replaced two 7950's for me just last week without issue. they both caught fire as well. They didn't ask any questions. I just submitted the RMA form and explanation that the cards shorted out and there was burn damage. I was actually running them slightly undervolted at 1070mV and 1000mhz core clock. I was not home when the cards caught fire. I had a big box fan blowing onto them though. They were installed in two different computers with 4x 7950's each. All the other msi 7950's were okay. The two cards that burned up had sequential serial numbers. Another card 2 serial numbers later also went defective (no post), but no burn damage. MSI isn't giving me any grief. I'm super glad because I bought a lot of these cards. I will definitely purchase MSI in the future.

Here's the pictures of my two fried 7950's:
SAM_1395.jpg

SAM_1402.jpg
Click to expand...
 
J Macker said:
I had the exact same issue with 2 msi 7950's.
It happened in almost the same location.

it looked like the fault was the underside of my cards at the 3rd capacitor from the bottom of the card.
This occurred when I did not have auto-gpu turned on. Since then I have turned auto-gpu on and haven't had any problems. I have 2 computers with 4x7950's each and 2 comps with 3x7950's each and I'm using PCP&C 950w psus in each comp. Yes, a 950w psu powering the 4x7950's. However, I have my cards undervolted to 1030 to 1050mV (reduced voltage a little more since this incident).

The 7950's that fried on me had dual 6-pin connectors. All my 8pin cards are fine.
I love MSI. They replaced the cards without giving me any problems. I didn't have to make a thread to turn some heads in order to get MSI to stand behind their warranty. It just took about 8-10 days before they shipped replacements.

Did you have auto-gpu turned on? What core/mem speeds? voltages? temps?
I've set temp cutoff at 93C now, target 84C. Using my IR temp laser while mining I found this area on the back of the PCB is the hottest part of the gpu.
Click to expand...

Hmm..

All of these cards are 6/8 pins.
voltage: 1.090V (stock is fixed at 1.250V but i flashed to an older bios version that runs lower)
core/mem: 1000/1250 (stock)
temps: ~75-80C stable at load, 90C cutoff

auto-gpu isn't turned on. I'll try running with that.

I don't know if the problem is occuring during power on or if the card failed while it was running to cause the crash (which then catches on fire next power on)

.
 
Sarian said:
Hmm..

All of these cards are 6/8 pins.
voltage: 1.090V (stock is fixed at 1.250V but i flashed to an older bios version that runs lower)
core/mem: 1000/1250 (stock)
temps: ~75-80C stable at load, 90C cutoff
auto-gpu isn't turned on. I'll try running with that.

.
Click to expand...

Auto-gpu will only throttle your cards when they get above target temp (or overheat-temp, can't remember)

Your numbers all look good, I'm using 1000mhz as well. Your temps are way better than mine, probably because of the hot ambient temps I have. It doesn't make sense to me why we had these cards burn in pretty much the same locations. It could just be faulty components.
 
yup, just not as durable as the 69XX cards were, I flogged 3 of those for 2.5 years mining 24/7 and never had an issue
 
J Macker said:
Auto-gpu will only throttle your cards when they get above target temp (or overheat-temp, can't remember)

Your numbers all look good, I'm using 1000mhz as well. Your temps are way better than mine, probably because of the hot ambient temps I have. It doesn't make sense to me why we had these cards burn in pretty much the same locations. It could just be faulty components.
Click to expand...

I put together an open setup in a basement that stays around 60-70F year round. It works out pretty well :D

I guess the 7950's aren't tough enough.
 
Last edited:
jojo69 said:
yup, just not as durable as the 69XX cards were, I flogged 3 of those for 2.5 years mining 24/7 and never had an issue
Click to expand...

I started mining back in 2011 with some sapphire 6950's and powercolor 6950's... Those things are still mining today and never had a single issue or instability.

I did have to replace the fans on some of them though but powercolor was awesome about that.
 
Unknown-One said:
I get that we're all primed to rip on Asus, but how long are we going to keep citing that thread? lol
Click to expand...

That's not the first or last time it's going to happen, as witnessed by numerous reports of similar behavior on the part of ASUS.
If we as a community don't give them shit where shit is due, they will continue to behave this way.

I have no problem bringing this thread up if it's relevant.;)

I also have no report from Derpentine that the problem has been resolved.....although it looks like it will be.

Look, people need to know what they're in for with ASUS if this happens and isn't their fault.
 
He's lucky they didn't just call the card "End of Life" and tell him to shove off. That's basically what they did with me and my Asus motherboard (which isn't even 2 years old).

No new drivers, no new BIOS... same shit I dealt with when I bought Foxconn motherboards, and was hoping to avoid.

OP is dealing with Gigabyte, though. I've had nothing but good luck with their RMA department, the few times I've had to use it. Not uncommon for them to replace your card with a higher-end version when you experience multiple failures like this.
 
You must log in or register to reply here.
Back
Top