Question MSI R9 280x Shutting off after some load

[panferno27]

Hello my friends! I have got a MSI R9 280x
I used bluestacks (clash of clans) and watched a video (at the same time) for like 15-20 minutes and suddenly my PC froze and I had to shut my pc down. Im quite sure that the heavy load (temperature problem?) caused my gpu to shut down. After that I couldnt start my PC it would give no signal (no beep, normally it would beep) but when I wait some time it would work again and fail again under load. Im quite sure that the GPU is the problem because I have another r9 280x at home which works fine under the same conditions without freezing once in days. I replaced the thermal paste. In the pictures below you can see my gpu specs. What i found suspicious is that the two metal things / spikes (whats the name of them) right under a R47 inductor chip (right bottom of the 2nd link) are bent and one of them even is slightly brownish. Could they be the cause for the failure and if yes can you fix it? Thank you for any answers would really help me out

https://gpuz.techpowerup.com/22/02/12/myn.png

http://puu.sh/III35/72f831a244.jpg

System: acer predator g5910 i7 2600k 16 gb ddr3 2tb

 

[Leeea]

no beep, normally it would beep

it is unlikely this is a video card issue


While different mainboards have different video card beep codes, they all have video card failure beeps.


Your issue sounds like a power supply failure. No beeps usually equals no power.

How old is the power supply?

because I have another r9 280x at home

In the same* computer or a different computer?

*(as in, you physically swapped it in)

 

[panferno27]

it is unlikely this is a video card issue


While different mainboards have different video card beep codes, they all have video card failure beeps.


Your issue sounds like a power supply failure. No beeps usually equals no power.


In the same computer or a different computer?

Hi I used all of them in the same System. Before the MSI r9 280x i had an ASUS r9 280x direct cu II in my pc (https://www.techpowerup.com/gpu-specs/asus-r9-280x-directcu-ii-top.b2441) and it never caused any problems.

 

[Leeea]

What power supply and how old?

 

[Leeea]

What i found suspicious is that the two metal things / spikes (whats the name of them) right under a R47 inductor chip (right bottom of the 2nd link) are bent and one of them even is slightly brownish. Could they be the cause for the failure and if yes can you fix it?

A TV repair shop or an electronics repair shop can replace it without to much difficulty.

However, 9 out of 10 times the reason that part overheats is power supply failure. (or one of the wires going from the power supply to the board) Specifically, the voltage starts to drop, the amps go up, and the component over heats.

You do have 2 separate wires running from the power supply to the GPU? One for each connector?

People on the cheap side of things have been known to just use thermal epoxy to attach a larger heatsink to components like that. Sometimes that works.


However, as previously mentioned, with a dying power supply futzing with the video card is just a waste of time.

 

[panferno27]

A TV repair shop or an electronics repair shop can replace it without to much difficulty.

However, 9 out of 10 times the reason that part overheats is power supply failure. Specifically, the voltage starts to drop, the amps go up, and the component over heats.

But how if I never had such a similar problem but right when i put this gpu in after some minutes it causes my pc to shut down and wont even start up again after some while.

 

[panferno27]

What power supply and how old?

What power supply and how old?

i dont know the name because i bought this pc when i was a kid 10 years ago

 

[Leeea]

and wont even start up again after some while

The power supply has a thermal breaker.

When power supplies fail or the power supply exceeds its specification the breaker overheats (trips), and must cool down before it will reset. No power is provided until it resets.

This is the exact symptom you are indicating you have.


What power supply and how old?

 

[panferno27]

The power supply has a thermal breaker.

When power supplies fail or the power supply exceeds its specification the breaker overheats (trips), and must cool down before it will reset. No power is provided until it resets.

This is the exact symptom you are indicating you have.


What power supply and how old?

750W and 10 years old doesnt know the name https://www.acer.com/datasheets/2012/4871/PredatorG5910/DT.SFJEG.008.html

 

[Leeea]

750W and 10 years old doesnt know the name https://www.acer.com/datasheets/2012/4871/PredatorG5910/DT.SFJEG.008.html

interesting

You are right at the point both parts would age out ( the GPU and the power supply )


Did the ASUS gpu come with the computer?

Are you able to try the MSI in a different computer?


What made you decide to swap them? You are indicating they are both 280x?

 

[panferno27]

interesting

You are right at the point both parts would age out ( the GPU and the power supply )

Did the ASUS gpu come with the computer?

Are you able to try the MSI in a different computer?

What made you decide to swap them?

no i bought both gpus recently after my old gpu (hd 7970) died. the one was for my friend who built a new pc and doesnt want to buy a new one cuz of the prices. im sadly not being able to try it on a different pc. But theres one thing 3 hrs ago i changed the thermal paste and since then it didnt crash once. Does that happen? Was it the cause?

 

[Leeea]

But theres one thing 3 hrs ago i changed the thermal paste and since then it didnt crash once. Does that happen? Was it the cause?

yes

when parts heat up, they require more power to continue running. This causes them to generate more heat, causing them to require even more power. With power supplies that are close to the edge of their capability, this will trigger the previously mentioned thermal breaker.


Sometimes with the wires that go to the GPU to the power supply get old, they develop little connection issues. Specifically right at the connector with the cheap generic ones like the ones you have. Simply unplugging them and plugging them back in moves them around a bit, and can fix the issue for a time.


It is great you managed to temporarily solve things. But this problem is likely to be back before the year is out.

Good luck.

750W and 10 years old doesnt know the name https://www.acer.com/datasheets/2012/4871/PredatorG5910/DT.SFJEG.008.html

These types of OEM units do not have a reputation for reliability. The cheap Chinese caps begin to degrade after 5 years. Most common symptoms of failure are freezing, spontaneous restarts, and spontaneous shutdown and wait.


When you do buy a replacement, be sure to get a brand name unit from a reputable seller. Lots of power supply scams / cheap crap with fake stickers on ebay these days.

 

[panferno27]

yes

when parts heat up, they require more power to continue running. This causes them to generate more heat, causing them to require even more power. With power supplies that are close to the edge of their capability, this will trigger the the previously mentioned thermal breaker.


Sometimes with the wires that go to the GPU to the power supply get old, they develop little connection issues. Specifically right at the connector with the cheap generic ones like the ones you have. Simply unplugging them and plugging them back in moves them around a bit, and can fix the issue for a time.


It is great you managed to temporarily solve things. But this problem is likely to be back before the year is out.

Good luck.


These types of OEM units do not have a reputation for reliability.


When you do buy a replacement, be sure to get a unit from a reputable seller. Lots of power supply scams / cheap crap with fake stickers on ebay these days.

ok thanks for your answer. Ill see

 

[Leeea]

ok thanks for your answer. Ill see

if you are going to let things be, one thing you can do is pick up a ir thermometer.

While under load, point it at the power regulator you were suspicious of. If it is over 120 Celsius you might want to consider giving it a bit of extra cooling.


The other test is the finger burn test. If it leaves a burn on your finger, it is to hot.