Question Is it my semi old power supply causing my PC to crash during games?

[ComputerPastaSweat]

Hi, I built my PC when I was teenager 8 years ago and ive always had crashing issues and I always thought it was a cooling problem so I upgraded my cooling but they still persist.
I ran some stress tests with BurnInTest of the RAM, GPU and CPU at 100% and it didn't crash and no errors were detected and temperatures were normal this was done for 30 minutes. The PerformanceTest said that my RAMs Memory Mark was 1252 which is in the low tier section, is this normal for my RAM? Here is a user benchmark test https://www.userbenchmark.com/UserRun/26085620

I can play non demanding games like CSGO for hours no problem but I can't play Rise of the graphical intensive games for less than 10minutes without my PC going to a black screen and restarting. I get around 60-90 fps on high settings in these games and they runs smoothly no issues but suddenly it just crashes and my pc restarts. This occurs for other demanding games sometimes like The Witcher 3, Battlefield 1, Rise of the Tomb Raider. I have monitored my temperatures and they are all below 60C. No crash reports nothing.

All drivers and BIOS are updated to the latest version.
This has stopped me from purchasing games that I want to play because Im scared that it cant run it for more than a gaming session worth of hours and its making me sad. Here is a YT video of my PC crashing:

View: https://www.youtube.com/watch?v=j5YOBe-piGc&feature=youtu.be


I have reseated everything apart from the CPU, disconnected all cables and spray them with air to clear dust.

Thanks for your help.

System:
Corsair Builder Series CXM 750W
MSI 970 Gaming Mother Board
1x 8GB DDR3 RAM Corsair CML8GX3M1A1600C10 Vengeance Low Profile 8GB (1x8GB) 1600 Mhz
AMD FX-8320 Black Edition 3.5GHz
NVIDIA GTX 970 4GB
NZXT Kraken x61 280mm
Seagate Barracuda 2TB
Samsung 250 GB SSD

 

[Falkentyne]

Yes it is entirely possible.
But before you run off and replace the PSU, contact Corsair and see if they can send you replacement PCIE cables for that exact PSU. Cables can also degrade and drop the GPU +12v signal and it drops too low--black screen. And this can happen even if the PSU is perfectly fine and putting out proper volts.

 

[jonnyguru]

Make sure "automatically restart" is unchecked:

https://support.microsoft.com/en-us...restarts-problem-when-windows-7-experiences-a

(It says "Windows 7", but this is also in Windows 8 and 10.)

 

[ComputerPastaSweat]

Yes it is entirely possible.
But before you run off and replace the PSU, contact Corsair and see if they can send you replacement PCIE cables for that exact PSU. Cables can also degrade and drop the GPU +12v signal and it drops too low--black screen. And this can happen even if the PSU is perfectly fine and putting out proper volts.

I called Corsair and my 5 year warranty is over sadly they said they couldn't help I wish I could go to my mates house and borrow his PSU but lockdown and all. Thanks for replying

 

[ComputerPastaSweat]

Make sure "automatically restart" is unchecked:

https://support.microsoft.com/en-us...restarts-problem-when-windows-7-experiences-a

(It says "Windows 7", but this is also in Windows 8 and 10.)

Hello, ok I will try this. Is this dangerous though?

 

[ComputerPastaSweat]

Make sure "automatically restart" is unchecked:

https://support.microsoft.com/en-us...restarts-problem-when-windows-7-experiences-a

(It says "Windows 7", but this is also in Windows 8 and 10.)

So I disabled automatically restart and I was able to use my computer for a lot longer than normal but it still crashed and restarted eventually. The Event Viewer showed that it was a Kernel-Power failure... what do you think? Bad PSU? other components were below 60C at 100% load. Thanks

 

[jonnyguru]

So I disabled automatically restart and I was able to use my computer for a lot longer than normal but it still crashed and restarted eventually. The Event Viewer showed that it was a Kernel-Power failure... what do you think? Bad PSU? other components were below 60C at 100% load. Thanks

Possibly.

 

[Vic 40]

Possibly.

Could it be about which gtx 970? (which gtx 970 is also a question to OP) Some seemed to created power spikes if i've remembered well (read about that) and maybe the psu can't cope with that. From what you told do those older type psu's have problems with that.

 

[Falkentyne]

I called Corsair and my 5 year warranty is over sadly they said they couldn't help I wish I could go to my mates house and borrow his PSU but lockdown and all. Thanks for replying

Did your PSU come with only two PCIE cables? Or only one? I just looked up a review and it says it comes with two 6+2 pin split cables (so 2 cables, 4 connectors--although you should use 1 card per cable).

I just looked up your video card.
The 970 4 GB only has one PCIE connector on it.
So in that case, why don't you remove the PCIE connector and use the spare one in your PSU box, and see if that helps ?

One good thing (which people overlook) about higher wattage power supplies (like the OneSeasonic initiative PSUs) is that you get multiple PCIE cables. If a cable degrades (You can see the GPU 12v monitoring on an AMD R9 290X reference card; Vega 56/64 won't show it, although it's possible some custom non reference cards will show it--), it can make you crash. That's exactly what happened to me when I used my Seasonic Platinum 1000W with a 2 connector 6/8 pin which plugged into one PSU port. The cable degraded over time and the 11.75v dropped over years to 11.25 and the black screens started. It came with 3 sets of split 6/8 cables (I didnt know at the time I'm supposed to use ONE cable per video card connector!), so when I threw away the degraded cable and used the last 2 cables, the 12v was now a nice flat 11.75v at idle and 11.63v at full load (this is what the original single cable read when I first bought my r9 290x) and all the black screen crashes stopped.

So in my case, the PSU was completely fine. The original cable was not. That's why I was mentioning that to you.

Now that same PSU is powering my Vega 64. But I do have a Seasonic PX-1000 sitting around in a new testbed experimental system which I haven't hooked up to my main system yet; will have to see if overclocking is improved with it soon.

BTW, Corsair is a pretty nice company.
Call them again and see if you can reach their level 2 tech support team on the phone and ask them very nicely if you can buy replacement PCIE cables from them. If they have them available, they can sell them to you even if it's not under warranty anymore. The big thing is if they have them available. Usually spare parts are only kept around for products that are not end of life.

Are these the exact cables? Because you should have another one available to swap out already.

https://www.ebay.com/i/113717943552

 

[ComputerPastaSweat]

Could it be about which gtx 970? (which gtx 970 is also a question to OP) Some seemed to created power spikes if i've remembered well (read about that) and maybe the psu can't cope with that. From what you told do those older type psu's have problems with that.

I have a MSI GTX 970 Million Edition. Ill do some more research. What solutions are they? just to replace?

 

[ComputerPastaSweat]

Did your PSU come with only two PCIE cables? Or only one? I just looked up a review and it says it comes with two 6+2 pin split cables (so 2 cables, 4 connectors--although you should use 1 card per cable).

I just looked up your video card.
The 970 4 GB only has one PCIE connector on it.
So in that case, why don't you remove the PCIE connector and use the spare one in your PSU box, and see if that helps ?

One good thing (which people overlook) about higher wattage power supplies (like the OneSeasonic initiative PSUs) is that you get multiple PCIE cables. If a cable degrades (You can see the GPU 12v monitoring on an AMD R9 290X reference card; Vega 56/64 won't show it, although it's possible some custom non reference cards will show it--), it can make you crash. That's exactly what happened to me when I used my Seasonic Platinum 1000W with a 2 connector 6/8 pin which plugged into one PSU port. The cable degraded over time and the 11.75v dropped over years to 11.25 and the black screens started. It came with 3 sets of split 6/8 cables (I didnt know at the time I'm supposed to use ONE cable per video card connector!), so when I threw away the degraded cable and used the last 2 cables, the 12v was now a nice flat 11.75v at idle and 11.63v at full load (this is what the original single cable read when I first bought my r9 290x) and all the black screen crashes stopped.

So in my case, the PSU was completely fine. The original cable was not. That's why I was mentioning that to you.

Now that same PSU is powering my Vega 64. But I do have a Seasonic PX-1000 sitting around in a new testbed experimental system which I haven't hooked up to my main system yet; will have to see if overclocking is improved with it soon.

BTW, Corsair is a pretty nice company.
Call them again and see if you can reach their level 2 tech support team on the phone and ask them very nicely if you can buy replacement PCIE cables from them. If they have them available, they can sell them to you even if it's not under warranty anymore. The big thing is if they have them available. Usually spare parts are only kept around for products that are not end of life.

Are these the exact cables? Because you should have another one available to swap out already.

https://www.ebay.com/i/113717943552

Im currently at uni and cant go home due to lockdown but that is where my spare cables are. Yes that eBay link are the cables that I use. Very interesting and lots for me to research and test. I might just buy some new ones instead of waiting to go home could be months. Thanks for replying and helping.

 

[Vic 40]

I have a MSI GTX 970 Million Edition. Ill do some more research. What solutions are they? just to replace?

If to replace something would i look at psu.

If you could test with either other psu or gpu would that maybe give some info.

 

[ComputerPastaSweat]

I have done some more research and it turns out that my GPU GTX 970 100ME requires a PCIE 3.0x16 and my Mobo only supports 2.0x16. Could this cause crashing issues?

If to replace something would i look at psu.

If you could test with either other psu or gpu would that maybe give some info.

 

[Vic 40]

No that is fine.