Idle & Load Overheating Sapphire R9 280x Tri-X 3GB

[nicktnack]

I have the Acer Predator G3-605 and I've recently decided to upgrade some parts, namely my PSU and GPU. Freshly installed, all new. I get the black screen of death and sound loop while playing games.

PSU Model: Corsair VS650
GPU: Sapphire R9 280x Tri-X 3GB

Idle Temps: 50c~60c
Full Load (FarCry 4 on Ultra): 94c~100c

I'm using TRIXX and these are my settings:
GPU Clock: 1000
Memory Clock: 1425
VDDC: 1131

Fan control is set at a strong curve once the GPU hits high temperatures, in this case, 60c.

Case is a little small and awkward hence the messy cables but I'm not sure that is the main problem here. Any suggestions? Is it a case of having to overhaul the case and installing more fans? Thanks.

 

[Anonymouselite5]

I have the Acer Predator G3-605 and I've recently decided to upgrade some parts, namely my PSU and GPU. Freshly installed, all new. I get the black screen of death and sound loop while playing games.

PSU Model: Corsair VS650
GPU: Sapphire R9 280x Tri-X 3GB

Idle Temps: 50c~60c
Full Load (FarCry 4 on Ultra): 94c~100c

I'm using TRIXX and these are my settings:
GPU Clock: 1000
Memory Clock: 1425
VDDC: 1131

Fan control is set at a strong curve once the GPU hits high temperatures, in this case, 60c.

Case is a little small and awkward hence the messy cables but I'm not sure that is the main problem here. Any suggestions? Is it a case of having to overhaul the case and installing more fans? Thanks.

Idle Temps: 50c~60c
Full Load (FarCry 4 on Ultra): 94c~100c

Are these temps your gpu temps or CPU temps?

Either way at 100 degrees on either of your parts your PC will shutdown to protect itself. Especially the CPU. If it's your cpu hitting these temps you should invest in an after market cooler for you CPU.

I suggest the Cooler master hyper EVO 212, just make sure your case is large enough to fit it in your case.
It's the best budget cpu cooler out there.

If you can spend on anything else, I'd say a new case or some case fans to push some of that hot air out and cool air in.


Another reason for your crashes could be your New gpu. There have been some problems on the r9 280x's with the memory clock being too high. In trixx try putting your memory clock from 1425 to 1300mhz. Many people have complained about crashing due to the memory.

Try pushing the memory down and post back here your results.

-good luck 😉

 

[nicktnack]

Those are GPU temps. So I just tried removing all the drivers with driver sweeper, CCleaner etc and reinstalled with the latest drivers.

Idle Temp is now ~40c

What else can I try?

 

[Anonymouselite5]

Those are GPU temps. So I just tried removing all the drivers with driver sweeper, CCleaner etc and now it won't detect my 280x. ATI's auto detect utility is also only detecting my onboard GPU.

GPU is running, lights, fans, the works. HDMI is plugged into GPU and I'm getting display on my monitor but device manager only shows my onboard gpu.

Don't use the auto detect utility. It's terrible.

Manually download the AMD driver from their website and install it. Restart and see if your gpu turns up again. Your gpu is still connected since the display still does rout through your card to your monitor.

As for your temps, if you run with your case open do they improve? If they don't it could be thermal problems with the gpu's cooler.

 

[nicktnack]

I realized CPU-Z detected it as 7970 so I installed the drivers anyway, I'm running FC4 now and the temp is about 74-75. Is that still too hot?

EDIT: Btw, this is with the CPU casing open. 20 mins later and it's 80C +

EDIT 2: Oops, before I typed this post. My PC did a sudden shut off at 85C. I'm at a loss.

 

[Anonymouselite5]

I realized CPU-Z detected it as 7970 so I installed the drivers anyway, I'm running FC4 now and the temp is about 74-75. Is that still too hot?

EDIT: Btw, this is with the CPU casing open. 20 mins later and it's 80C +

EDIT 2: Oops, before I typed this post. My PC did a sudden shut off at 85C. I'm at a loss.

Have you tried downclocking the memory clock like before? So it doesn't crash.
80 degrees is the barrier, if it goes above that then you have a problem.

In trixx fix the fans to 100% and play a game, and record the highest temp you get (with the case open).
That's quite a nice card, I have the r9 290 version of it, and the r9 290 gets much hotter than the r9 280x (7970). It usually never breaks 72 degrees.
So yours is too hot.

If your temps are way lower with the fans at 100 and the case open then you may need better airflow into your case and perhaps a new case or casefans are needed.

-good luck

 

[nicktnack]

Hello, I downclocked to 1300 and tried FC4 and Shadow of Mordor. I'm capping around 70C~80C with the case closed, no sudden freezes or shutdowns occurring yet. Idle is around 50C~60C

 

[Anonymouselite5]

Hello, I downclocked to 1300 and tried FC4 and Shadow of Mordor. I'm capping around 70C~80C with the case closed, no sudden freezes or shutdowns occurring yet. Idle is around 50C~60C

So far your crashes seem to have been fixed?
Your temps are still too high for an r9 280x, what you can try (if you feel confident). A few other people fixed their temp issues this way.
Buy some thermal paste and take off the cards cooler, and clean off the previous thermal paste and reapply your own.
(note this can void your warranty, but if you don't make a mistake they can't tell) A few other people brought their temps down to about 60 under load after doing this on the Toxic r9 280x.
If you don't want to do it, you could return it in exchange for a replacement.

-good luck

 

[nicktnack]

Hello, I downclocked to 1300 and tried FC4 and Shadow of Mordor. I'm capping around 70C~80C with the case closed, no sudden freezes or shutdowns occurring yet. Idle is around 50C~60C

So far your crashes seem to have been fixed?
Your temps are still too high for an r9 280x, what you can try (if you feel confident). A few other people fixed their temp issues this way.
Buy some thermal paste and take off the cards cooler, and clean off the previous thermal paste and reapply your own.
(note this can void your warranty, but if you don't make a mistake they can't tell) A few other people brought their temps down to about 60 under load after doing this on the Toxic r9 280x.
If you don't want to do it, you could return it in exchange for a replacement.

-good luck

It unfortunately crashed again, idle temps were up to 65C in an AC room. I've emailed Sapphire's support to see if I can get a replacement before trying the thermal paste. Thanks for the help. I'll update here again soon.

EDIT: Just wondering, would this be a viable solution?
http://www.techpowerup.com/forums/threads/solution-is-found-to-sapphire-vapor-x-r9-280x-vrm-over-heating-and-throttling-problem.196972/

 

[Anonymouselite5]

You can try updating or going back to an older bios, just make sure you keep a backup of the original bios so if you do try to RMA it (replace it) they will accept it.
Note: the card also has a dual bios switch on it.

Also if you brought the card from a store go there first before Sapphire, even if the Bios fixes the temps. It may not fix the crashing on it.
Also one more attempt to try stop crashing.
Go to Trixx and set the power limit to +20%, set the memory to 1200mhz, and the core clock to 800mhz.
If it still crashes it is most likely a very faulty bios or a very faulty card.

-good luck

 

[nicktnack]

You can try updating or going back to an older bios, just make sure you keep a backup of the original bios so if you do try to RMA it (replace it) they will accept it.
Note: the card also has a dual bios switch on it.

Also if you brought the card from a store go there first before Sapphire, even if the Bios fixes the temps. It may not fix the crashing on it.
Also one more attempt to try stop crashing.
Go to Trixx and set the power limit to +20%, set the memory to 1200mhz, and the core clock to 800mhz.
If it still crashes it is most likely a very faulty bios or a very faulty card.

-good luck

What does the dual bios switch do if I activate it? Trixx doesn't allow me to set the power limit I think and MSI Afterburner does not allow me control the voltage (option is greyed out despite every fix I've tried)

Sapphire's Answer: Dear Customer :
The card is quiet new , you can try it with another system to isolate the problem at first.
If you don't have , please contact your retailer to check if the card defective or not ?
And the R9 280X series should need about 750W PSU , you can try to update the PSU and try again at first.
Thanks.

 

[Anonymouselite5]

You can try updating or going back to an older bios, just make sure you keep a backup of the original bios so if you do try to RMA it (replace it) they will accept it.
Note: the card also has a dual bios switch on it.

Also if you brought the card from a store go there first before Sapphire, even if the Bios fixes the temps. It may not fix the crashing on it.
Also one more attempt to try stop crashing.
Go to Trixx and set the power limit to +20%, set the memory to 1200mhz, and the core clock to 800mhz.
If it still crashes it is most likely a very faulty bios or a very faulty card.

-good luck

What does the dual bios switch do if I activate it? Trixx doesn't allow me to set the power limit I think and MSI Afterburner does not allow me control the voltage (option is greyed out despite every fix I've tried)

Sapphire's Answer: Dear Customer :
The card is quiet new , you can try it with another system to isolate the problem at first.
If you don't have , please contact your retailer to check if the card defective or not ?
And the R9 280X series should need about 750W PSU , you can try to update the PSU and try again at first.
Thanks.

Your psu is quite a cheap psu but it still should handle it. You don't need a 750w psu for an r9 280x.
Im using a 600w psu with an r9 290, this guy has a problem.
Even if your psu was bad it still doesn't answer why it overheats.

By the way it's not voltage control, it's the power limit. It's easier to see if you go to your catalyst control center and go to overdrive settings. Although like he said you can also try the gpu in another persons system, maybe a friend or a system at a PC store if they let you.

-good luck

 

[nicktnack]

You can try updating or going back to an older bios, just make sure you keep a backup of the original bios so if you do try to RMA it (replace it) they will accept it.
Note: the card also has a dual bios switch on it.

Also if you brought the card from a store go there first before Sapphire, even if the Bios fixes the temps. It may not fix the crashing on it.
Also one more attempt to try stop crashing.
Go to Trixx and set the power limit to +20%, set the memory to 1200mhz, and the core clock to 800mhz.
If it still crashes it is most likely a very faulty bios or a very faulty card.

-good luck

What does the dual bios switch do if I activate it? Trixx doesn't allow me to set the power limit I think and MSI Afterburner does not allow me control the voltage (option is greyed out despite every fix I've tried)

Sapphire's Answer: Dear Customer :
The card is quiet new , you can try it with another system to isolate the problem at first.
If you don't have , please contact your retailer to check if the card defective or not ?
And the R9 280X series should need about 750W PSU , you can try to update the PSU and try again at first.
Thanks.

Your psu is quite a cheap psu but it still should handle it. You don't need a 750w psu for an r9 280x.
Im using a 600w psu with an r9 290, this guy has a problem.
Even if your psu was bad it still doesn't answer why it overheats.

By the way it's not voltage control, it's the power limit. It's easier to see if you go to your catalyst control center and go to overdrive settings. Although like he said you can also try the gpu in another persons system, maybe a friend or a system at a PC store if they let you.

-good luck

Uh oh, looks the bios downgrade didn't go well. Now loads black screen when I restart. I have to activate the alternate bios through the GPU to load up.

What do I do now?

 

[Anonymouselite5]

You'll have to flash the other side of the switch back to the original bios.

Here is one way:
http://www.overclock.net/t/593427/how-to-unbrick-your-bricked-graphics-card-fix-a-failed-bios-flash

I believe there is an easier way since you have the dual bios, but I've never actually needed to use it 😉
Once you've done that, at this rate you may consider RMA'ing the card.

 

[nicktnack]

You'll have to flash the other side of the switch back to the original bios.

Here is one way:
http://www.overclock.net/t/593427/how-to-unbrick-your-bricked-graphics-card-fix-a-failed-bios-flash

I believe there is an easier way since you have the dual bios, but I've never actually needed to use it 😉
Once you've done that, at this rate you may consider RMA'ing the card.

Restored the bricked BIOS. What's next? RMA? I've tried lowering the numbers as per your advice but still pretty much the same I guess. I can try again. I'm actually not sure if I should try flashing, I just checked and it seems like I have the latest BIOS version anyway.

 

[Anonymouselite5]

Yes I would RMA. It overheats, it crashes your PC. You've tried lowering clocks and upping fan speed.
You can try updating your driver but if you already have the AMD omega driver (it's pretty stable for me) then I would just get a replacement.

-good luck

 

[nicktnack]

Yes I would RMA. It overheats, it crashes your PC. You've tried lowering clocks and upping fan speed.
You can try updating your driver but if you already have the AMD omega driver (it's pretty stable for me) then I would just get a replacement.

-good luck

I put back my 3-4 years old 5850 HD and it's idling at 90C+. Super ridiculous, but I'm not getting crashes. Now I'm really not sure what's wrong here.. Tried adjusting the fan speed in Overdrive but it didn't increase too. Weird.. FYI, I didn't change the drivers.

 

[Anonymouselite5]

Yes I would RMA. It overheats, it crashes your PC. You've tried lowering clocks and upping fan speed.
You can try updating your driver but if you already have the AMD omega driver (it's pretty stable for me) then I would just get a replacement.

-good luck

I put back my 3-4 years old 5850 HD and it's idling at 90C+. Super ridiculous, but I'm not getting crashes. Now I'm really not sure what's wrong here.. Tried adjusting the fan speed in Overdrive but it didn't increase too. Weird.. FYI, I didn't change the drivers.

Idling at 90 degrees? Ive seen other peoples 5850's get to 95 degrees under load, but idle?
Do you run Trixx at startup? Go into trixx's settings and is the option disable ulps (or continuous voltage) ticked on?
If it is tick it off.
The crashing stopped but the heat? Something is wrong here, and it can't be your case since you get the same with it open. Is any of the voltage overvolting options been touched in the trixx settings?
Also download GPU-z if you don't have it,
http://www.techpowerup.com/gpuz/

Look at the voltage (VCCD), gpu load, core clock, fan speed, and any other info you think may be important. post them here. Also try putting back in your r9 280x and posting what it says in GPU-z for the 280x here as well.
Also if you haven't already tried, un-install your graphics driver and reinstall with the latest Version. after you've looked up in GPU-z.

-good luck

 

[nicktnack]

Yes I would RMA. It overheats, it crashes your PC. You've tried lowering clocks and upping fan speed.
You can try updating your driver but if you already have the AMD omega driver (it's pretty stable for me) then I would just get a replacement.

-good luck

I put back my 3-4 years old 5850 HD and it's idling at 90C+. Super ridiculous, but I'm not getting crashes. Now I'm really not sure what's wrong here.. Tried adjusting the fan speed in Overdrive but it didn't increase too. Weird.. FYI, I didn't change the drivers.

Idling at 90 degrees? Ive seen other peoples 5850's get to 95 degrees under load, but idle?
Do you run Trixx at startup? Go into trixx's settings and is the option disable ulps (or continuous voltage) ticked on?
If it is tick it off.
The crashing stopped but the heat? Something is wrong here, and it can't be your case since you get the same with it open. Is any of the voltage overvolting options been touched in the trixx settings?
Also download GPU-z if you don't have it,
http://www.techpowerup.com/gpuz/

Look at the voltage (VCCD), gpu load, core clock, fan speed, and any other info you think may be important. post them here. Also try putting back in your r9 280x and posting what it says in GPU-z for the 280x here as well.
Also if you haven't already tried, un-install your graphics driver and reinstall with the latest Version. after you've looked up in GPU-z.

-good luck

I turned off ULPS and continuous voltage, did a restart. Idle temps are still way too high.

 

[Anonymouselite5]

Was your old card like this when you first got it? the temps.
If it wasn't, uninstall your AMD driver.
download driver sweeper:
http://www.guru3d.com/content-page/guru3d-driver-sweeper.html

First uninstall through the control panel.
Boot back up in safe mode.
Then go to driver sweeper (while in safe mode) and uninstall your AMD driver through it.
Boot back into normal mode and download the driver again. Also possibly it could be the newest driver causing the problems, so if that doesn't fix anything you can google for the an older driver (possibly 14.9) and find it on guru's site.

Your 5850 clocks seem fine and voltage, What were the temps like before?

edit: Also try with your case open.

 

[nicktnack]

Was your old card like this when you first got it? the temps.
If it wasn't, uninstall your AMD driver.
download driver sweeper:
http://www.guru3d.com/content-page/guru3d-driver-sweeper.html

First uninstall through the control panel.
Boot back up in safe mode.
Then go to driver sweeper (while in safe mode) and uninstall your AMD driver through it.
Boot back into normal mode and download the driver again. Also possibly it could be the newest driver causing the problems, so if that doesn't fix anything you can google for the an older driver (possibly 14.9) and find it on guru's site.

Your 5850 clocks seem fine and voltage, What were the temps like before?

edit: Also try with your case open.

Welp. So I went about trying to uninstall the drivers and it went all black screen. I put the 280x back and am in process of installing with fresh ATI drivers (14.11.2 Beta, because what the heck since 14.12 wasn't any better anyway) after removing with Driver Sweeper and CCleaner.

EDIT: 280x is now idling at 40-50C, full load at 76C on FC4. This is without any voltage or clock throttling but I have a custom fan profile active.

 

[Anonymouselite5]

That's much better temp wise, Still getting crashes on it? If not it seems that your driver before was faulty.
Had to driver sweep like that a few times before with my cards as well, but usually it wasn't as bad, just a few crashes here and there.

Your 280x is exactly at normal 280x load temps, is your card fine now?

 

[nicktnack]

That's much better temp wise, Still getting crashes on it? If not it seems that your driver before was faulty.
Had to driver sweep like that a few times before with my cards as well, but usually it wasn't as bad, just a few crashes here and there.

Your 280x is exactly at normal 280x load temps, is your card fine now?

Everything looks good save for some FPS drops but I can live with that. No crashes in the past 2 hours with FC4 and Shadow of Mordor. Thanks for your patience!

 

[Anonymouselite5]

No problem, that's a very nice card 😉

 

[nicktnack]

Oh boy, multiple crashes today for some reason playing dragon age inquisition and dota 2. Temps peaked around 80C+ so I'm not sure why. Either way, I think I'm gonna do an RMA.