What is wrong with my PC?

[Grobai]

Hello Linustechtips. Recently I have been getting crashes whenever my computer is running something that requires power.

 

What I have experienced

• Whenever I play a game, the game crashes after a few minutes and my fans are making a lot of noise.
• I can play some games, League of Legends as an example, but not CS:GO, Fifa, for honor, or any other games.
• If I play League of Legends while watching a stream, it also crashes.

What I have tried to test the problem with

• I have used MSI Afterburner and MSI Kombuster to stress the GPU and when I do so, the test crashes when my GPU usage is at 100%, temperature is usually around 70-80 here.
• I have tried lowering the resolution in games like CS:GO and Fifa, and it works in CS:GO when I set my resolution to 600*400, but it doesn't work in Fifa.

What I have tried to resolve the issue

• Rebooting the PC.
• Cleaning the PC from any dust inside the case.
• Updating the driver.
• Playing without a driver.

Specs

• GPU: GTX 670
• CPU: i5-3570
• RAM: 8gb
• PSU: 550 Watt

 

My problem
I am not quite sure what my problem is, but I believe it to be my GPU based on my tests, but I don't have a deep knowledge or understanding of hardware so I am not sure. I have google a lot, and it seems that the PSU, RAM, and CPU could also be a problem. Any help is appreciated, as I need this to be fixed as fast as possible. If it is my GPU, what GPU would you recommend me getting? Been looking at 970 and 1070. Thanks in advance.

[TomR]

2 minutes ago, Grobai said:

Hello Linustechtips. Recently I have been getting crashes whenever my computer is running something that requires power.

 

What I have experienced

• Whenever I play a game, the game crashes after a few minutes and my fans are making a lot of noise.
• I can play some games, League of Legends as an example, but not CS:GO, Fifa, for honor, or any other games.
• If I play League of Legends while watching a stream, it also crashes.

What I have tried to test the problem with

• I have used MSI Afterburner and MSI Kombuster to stress the GPU and when I do so, the test crashes when my GPU usage is at 100%, temperature is usually around 70-80 here.
• I have tried lowering the resolution in games like CS:GO and Fifa, and it works in CS:GO when I set my resolution to 600*400, but it doesn't work in Fifa.

What I have tried to resolve the issue

• Rebooting the PC.
• Cleaning the PC from any dust inside the case.
• Updating the driver.
• Playing without a driver.

Specs

• GPU: GTX 670
• CPU: i5-3570
• RAM: 8gb
• PSU: 550 Watt

 

My problem
I am not quite sure what my problem is, but I believe it to be my GPU based on my tests, but I don't have a deep knowledge or understanding of hardware so I am not sure. I have google a lot, and it seems that the PSU, RAM, and CPU could also be a problem. Any help is appreciated, as I need this to be fixed as fast as possible. If it is my GPU, what GPU would you recommend me getting? Been looking at 970 and 1070. Thanks in advance.

What brand is the power supply?

[ThinkWithPortals]

First of all: thank you. Posts like this, with in-depth information, system specs and detail, make solving problems so much easier.

 

Now: what make is the power supply, specifically? And have you tried using the software Display Driver Uninstaller to do a complete clean uninstall and reinstall of GPU drivers?

[Dabombinable]

Its possible that the PSU is on its last legs, or if its a low uality one, not actually running at its rated voltage-I've got one for example that's rated at 400W, but its actually 250W combined across all 3 rails (the unit says so itself).

[Grobai]

5 minutes ago, ThinkWithPortals said:

First of all: thank you. Posts like this, with in-depth information, system specs and detail, make solving problems so much easier.

 

Now: what make is the power supply, specifically? And have you tried using the software Display Driver Uninstaller to do a complete clean uninstall and reinstall of GPU drivers?

I am using a TX650 Corsair PSU right now. And thank you, making it easier for you is making it easier for me.

[SaulkM]

Hello,

 

It may be a problem based on either your GPU or possibly your PSU. I doubt that your ram is the issue.

 

Purchasing a cheap PSU may have created this problem. If once the 670 gets to 100% load, your entire computer crashes, that may be it.

 

More likely, it may be your drivers need to be reinstalled or that the 670 may be getting tired.

 

If you feel a replacement is in order, I suggest a 970, based on the fact that the games you have listed are not the most graphically demanding. A 970 can be found at almost half the price as a 1070 nowadays.

[Dabombinable]

Just now, Grobai said:

I am using a TX650 Corsair PSU right now. And thank you, making it easier for you is making it easier for me.

I'd say that its faulty-all of the parts together would come any where near consuming even 400W.

 

http://www.anandtech.com/show/5818/nvidia-geforce-gtx-670-review-feat-evga/17

 

Edit: That test was with the far more power hungry i7 3960X @ 4.3GHz BTW

[TomR]

2 minutes ago, SaulkM said:

Hello,

 

It may be a problem based on either your GPU or possibly your PSU. I doubt that your ram is the issue.

 

Purchasing a cheap PSU may have created this problem. If once the 670 gets to 100% load, your entire computer crashes, that may be it.

 

More likely, it may be your drivers need to be reinstalled or that the 670 may be getting tired.

 

If you feel a replacement is in order, I suggest a 970, based on the fact that the games you have listed are not the most graphically demanding. A 970 can be found at almost half the price as a 1070 nowadays.

I haven't seen any 970s for sale for a reasonable price. The ones I'm seeing are going for new GTX 1060 prices.

[Grobai]

1 minute ago, Dabombinable said:

I'd say that its faulty-all of the parts together come any where near consuming even 400W.

I should probably also state that I have been using this setup for about 4 years with no problems, and it is only now that I have a problem.

[Megah3rtz]

5 minutes ago, Grobai said:

I am using a TX650 Corsair PSU right now. And thank you, making it easier for you is making it easier for me.

80+ or 80+ bronze or 80+ bronze v2 TX650

[Mello]

Do you know what part is heating, CPU or GPU?

and what fans that become noisy when you play games?

[Dabombinable]

Just now, Grobai said:

I should probably also state that I have been using this setup for about 4 years with no problems, and it is only now that I have a problem.

PSU can degrade over that span of time-though its definitely not common (and I've only had it happen once, with an inlcuded-with-case-PSU that's still kicking).

[Grobai]

1 minute ago, Mello said:

Do you know what part is heating, CPU or GPU?

and what fans that become noisy when you play games?

I do not know this. I only know that my GPU is around 65-85 degrees when I play games.

[doomsriker]

That 670 is pretty old. My guess is that your gpu's thermal paste has become dry and less effective. Try replaceing the thermal paste on your gpu. You can find guides on how to do this on YouTube. 

[ThinkWithPortals]

6 minutes ago, Grobai said:

I am using a TX650 Corsair PSU right now. And thank you, making it easier for you is making it easier for me.

It could be the problem, PSUs can wear out over time and start to play up like this. It would explain the fans spinning up really fast, if there are sporadic power surges... do you have another power supply you can test with?

[Grobai]

1 minute ago, ThinkWithPortals said:

It could be the problem, PSUs can wear out over time and start to play up like this. It would explain the fans spinning up really fast, if there are sporadic power surges... do you have another power supply you can test with?

No sadly I don't have another PSU to test with.

[Mello]

1 minute ago, doomsriker said:

That 670 is pretty old. My guess is that your gpu's thermal paste has become dry and less effective. Try replaceing the thermal paste on your gpu. You can find guides on how to do this on YouTube. 

Let's make sure what is causing the problem before getting into action, and maybe make more problems.

[Dabombinable]

Just now, doomsriker said:

That 670 is pretty old. My guess is that your gpu's thermal paste has become dry and less effective. Try replaceing the thermal paste on your gpu. You can find guides on how to do this on YouTube. 

Its not overheating though, and even with Kepler they can go to around 95^oC safely.

Just now, Grobai said:

No sadly I don't have another PSU to test with.

Having a look for reviews, it turns out that the TX650 has been around since 2008-and is made by Seasonic who are one of the good manufacturers, so if it is at fault its actually very unusual.

[SaulkM]

1 minute ago, TomR said:

I haven't seen any 970s for sale for a reasonable price. The ones I'm seeing are going for new GTX 1060 prices.

If you don't mind getting a re certified card, a 970 can be found on newegg for around the same price point as a cheap 3 gb 1060. It can be harder to get those older cards for a fair price.

[Grobai]

3 minutes ago, Dabombinable said:

Its not overheating though, and even with Kepler they can go to around 95^oC safely.

Having a look for reviews, it turns out that the TX650 has been around since 2008-and is made by Seasonic who are one of the good manufacturers, so if it is at fault its actually very unusual.

What would you recomend me doing then? I am absolutely out of ideas, and I don't want to go down and buy a 1070 and come home with a faulty PSU.

[doomsriker]

1 minute ago, Mello said:

Let's make sure what is causing the problem before getting into action, and maybe make more problems.

Replaceing GPU thermal paste is as likely to create problems as replaceing CPU thermal paste. That is to say not very likely. In my opinion, replaceing GPU thermal paste is a normal troubleshooting step as long as it does not void any warrenty. 

[Megah3rtz]

6 minutes ago, Dabombinable said:

Having a look for reviews, it turns out that the TX650 has been around since 2008-and is made by Seasonic who are one of the good manufacturers, so if it is at fault its actually very unusual.

No, not all are made by Seasonic. Some are made by CWT.

[TomR]

4 minutes ago, SaulkM said:

If you don't mind getting a re certified card, a 970 can be found on newegg for around the same price point as a cheap 3 gb 1060. It can be harder to get those older cards for a fair price.

Exactly my point. Why would you want a used 970 over a new (and better performing) 1060?

[Megah3rtz]

https://ca.pcpartpicker.com/product/mB3RsY/corsair-power-supply-cmpsu650tx Made by Seasonic

 

https://ca.pcpartpicker.com/product/vszv6h/corsair-power-supply-tx650 Made by CWT 

 

https://ca.pcpartpicker.com/product/Z4fp99/corsair-power-supply-cmpsu650txv2 Made by Seasonic

[SpaceGhostC2C]

17 minutes ago, Grobai said:

What would you recomend me doing then? I am absolutely out of ideas, and I don't want to go down and buy a 1070 and come home with a faulty PSU.

I think you may want to make sure you isolate the problem first. Fan noise could simply mean that the GPU is noisy when fans are cranked up, but if it stays at 85C max it shouldn't be GPU heat. Hence, I suggest monitoring CPU temps too and see how hot it gets under load.

 

Having said that, I think a good lead to follow is

Quote

I have used MSI Afterburner and MSI Kombuster to stress the GPU and when I do so, the test crashes when my GPU usage is at 100%, temperature is usually around 70-80 here.

Since that taxes the GPU only, we can focus on that and think of a way to separate heat issues from power issues. For example, using MSI Afterburner to monitor the GPU, what is the speed of the GPU fans during the stress tests? If it's below 100%, try setting it fixed at 100% (cover your ears ) and run another tests, to see if there is a difference in temperature and behavior (e.g., whether it manages to complete the test). You could still have other components of the graphics hard overheating (like VRMs) while the GPU itself is fine, but that would be harder to diagnose.

If temperatures don't seem to matter, you would like to make sure it is power delivery. That is harder to do without additional tools, but since your PSU has a single 12v rail, you could try the following: try to run a CPU stress test, so you load the 12v rail with a different component. If that fails as well, maybe the PSU no longer can deliver 12v stable under heavy load. If it doesn't, it still doesn't mean the PSU is fine (your CPU probably won't draw as much power as the GPU), opens up the possibility of the GPU being the problem.

 

Bottom line: for now, I think it is more accurate to say your computer crashes whenever requiring heavy GPU usage. If you try stressing the CPU, we could say more about whether it happens when requiring power in general or not