PC Shuts Down Without Warning During Gaming

Wargar_Bova

Commendable
Sep 28, 2016
8
0
1,510
Hello, I built my own rig from parts I've selected from PCPartPicker with hopes of not only playing the latest games but also doing a bit of video and audio editing as well. That being said after a week of setting it up and installing drivers, operating systems and the like I'm already experiencing issues.

I got the alpha version of Star Citizen to test out my graphics card and after a couple minutes of gameplay my computer would shut down. I tried this a few times at different resolutions and quality settings and I'd still experience the same issue. So I tried it with an already finished game, Fractured Space, ran it on high settings and experienced the same issue. My computer is shutting off after a couple minutes of gameplay and not restarting.

I've checked temperatures with Speedfan and Speccy, even bought a new fan just to keep the air flowing, got all three fans above 1000 RPM and yet the problem persists. I've racked my mind over temperatures that I've beginning to think it could be something else. Perhaps my power? I'll differ to your judgement.

My rig,

-Intel Core i5-6500 3.2GHz Quad-Core Processor
-Microsoft Windows 10 Home OEM 64-bit

-Corsair 200R ATX Mid Tower Case
-Cooler Master Hyper 212 EVO 82.9 CFM Sleeve Bearing CPU Cooler

-Asus Z170-P ATX LGA1151 Motherboard
-Corsair Vengeance LPX 16GB (2 x 8GB) DDR4-2666 RAM Memory
-Asus GeForce GTX 970 4GB STRIX Video Card
-EVGA 850W 80+ Bronze Certified Semi-Modular ATX Power Supply

-Sandisk SSD PLUS 120GB 2.5" Solid State Drive
-Seagate Barracuda 1TB 3.5" 7200RPM Internal Hard Drive



Thank you very much for your time,
 
Solution

The PSU does not need to go bad. PSUs have to cope with an extremely broad range of CPU and GPU load transients. It is entirely possible that the combination of PSU, GPU and CPU simply aren't playing nice together. Change any one variable in any way and everything may become perfectly fine.

Odd interoperability quirks may be frustrating but they do occur and I suspect they will become increasingly common as CPUs and GPUs push ever more aggressive power management, making the PSU's job that much more challenging.

Dusanking123

Reputable
Sep 2, 2014
63
0
4,640
Its possible that PSU is not strong enough,forcing your pc to shut down.That is because ur graphic card is too strong.If you want this graphic card get new PSU,if you cant afford it get older graphic card.Make choice fast,i didnt make choice and i couldnt run comp until i got old graphic card
 

Wargar_Bova

Commendable
Sep 28, 2016
8
0
1,510


Thank you, I'll have to look into it. It's just a pain in the wallet that I messed up on this.
 

Wargar_Bova

Commendable
Sep 28, 2016
8
0
1,510


Would you have any recommendations on a new power supply, that seams to be the more economic route.
 

Dusanking123

Reputable
Sep 2, 2014
63
0
4,640

I dont know which PSU would be good,i just know that this one is not good enough,try looking on net,if this helped select best solution :D
 

juanrdp

Honorable
Nov 7, 2012
857
0
11,360


A 850W not enought for a GTX 970?, you could even run a tri-SLI of 970s with 850w....
Another thing is that the PSU is failing, but you have power more than enought, the recommentded PSU for a 970 is only 500w and draw 150w.
 

InvalidError

Titan
Moderator
You checked the temperatures but didn't say what they were.

850W is vastly overkill for your PC and most EVGA PSUs are of at least decent quality, so you should be reasonably well covered in that department. A defective or problematic interaction between the PSU and its loads are still possibilities if you have ruled out all other likely causes.
 

juanrdp

Honorable
Nov 7, 2012
857
0
11,360


And what temperatures you get on the CPU and on the GPU while playing (just before the shutdown, use a layout program like Aida to get an idea)

 
EVGA 850W 80+ Bronze is a tier 2 unit and still good, but it doesn't go to say that it might still be a bad one that you got. The 970 only requires a 500w PSU so he has more then enough power.

What kind of temps were you getting? If you can try a different PSU from a friend. If its still within the 30 days try to send it back to who you bought it from and get a new one.
 

Wargar_Bova

Commendable
Sep 28, 2016
8
0
1,510
I'll have to report my exact SpeedFan and Speccy numbers tonight after I get off work, last I recorded though my CPU and cores stay stable around 20-30C and while running the Star Citizen alpha or Fractured Space my GPU will climb up to about 65C.

Again I'll have to report those numbers later today.
 

juanrdp

Honorable
Nov 7, 2012
857
0
11,360


Ok, i will try to check latter this night.
65º for a GTX 970 under load (and Star Citizen stress a lot the GPU) is a very good number, but 20º or even 30º for the CPU is waaay under the normal values, or you have the PC case inside a freezer or is no way that you could have a CPU core temp of 30º under load.

 

Wargar_Bova

Commendable
Sep 28, 2016
8
0
1,510
Thermal paste was very recently applied and I believe in correct fashion. At rest my CPU sits at 25C, GPU at 31C, Motherboard 28C, Harddrive and SSD on average at 26C.

I did stress tests on my CPU and GPU last night. Under stress my GPU got to about 70C, bounced around that for a few minutes and my system shut down. Before it shut down my CPU only reached 31C, my motherboard got up to 34C, and my SSD and harddrive didn't change much in temperature at all. Looked at my event log and no thermal errors popped up.

I ran a CPU test and got it up to 45C, bounced around that for a few minutes and nothing happened, so I can probably rule CPU out as the culprit.

 

nerro120

Honorable
Feb 13, 2014
418
0
10,810
So the computer doesn't restart at all? Until you do a hard reset I assume. What was the gpu load during your test and Do you have any other high power draw appliances or anything running to the same outlet
 

Wargar_Bova

Commendable
Sep 28, 2016
8
0
1,510



Nothing else is drawing power from the outlet, just the computer in question. It seems as though, and this is my observation, that I'm getting shut downs when my GPU is stressed. Weather it be via stress tests or high-graphic-quality gaming I can't seem to get more than 5 minutes out of it when the GPU gets above 65C.

 

nerro120

Honorable
Feb 13, 2014
418
0
10,810
Well 65 and 70 degrees should not cause a shut down unless it's power related, like we said before, it could be the psu causing these shutdowns. It sounds like to me atleast that it's something internal to the power supply since it only does it under load from the 970.
 

Wargar_Bova

Commendable
Sep 28, 2016
8
0
1,510


I wouldn't call myself an expert by any means, little experience as well so if you don't mind me asking how do I go about testing that? What should I look out for?

I'm okay with the idea of upgrading the power supply but I'd like to make sure I know what I need.

 

nerro120

Honorable
Feb 13, 2014
418
0
10,810
You did about the only testing that can be done by stressing the gpu to the point of shut down, besides replacing it and seeing if the new one works that's about it. The problem with it is that it's a pretty decent quality unit and way more than you would need and it doesn't make sense why it would do that but sometimes good stuff goes bad prematurely and it sucks but it's part of the game.
 

InvalidError

Titan
Moderator

The PSU does not need to go bad. PSUs have to cope with an extremely broad range of CPU and GPU load transients. It is entirely possible that the combination of PSU, GPU and CPU simply aren't playing nice together. Change any one variable in any way and everything may become perfectly fine.

Odd interoperability quirks may be frustrating but they do occur and I suspect they will become increasingly common as CPUs and GPUs push ever more aggressive power management, making the PSU's job that much more challenging.
 
Solution

nerro120

Honorable
Feb 13, 2014
418
0
10,810
I did not say it needed to go bad, I was just saying that the psu is not liking the load of the gpu or the gpu is not liking the power supply, without load tests and proper equipment to test the psu, it is the easiest option to replace and see if a new one works