PC is crashing while playing games

[Kuhulnim]

Hello,
My pc crashes at a random*** moment while I am playing games on it. It doesn't give any blue screen, it just shuts down completely and, 2 seconds later, it turns on again but nothing is displayed on my monitors and there's no beep sound. So I need to shut it down by holding the power button. If I try to turn it on again, the same thing happens, no beep and no display. In order to be able to really make it work, I need to wait at least 30 minutes.
When I'm doing normal stuff, there is absolutely no crashes. I can leave my pc turned on for one week and it won't crash.
I've also noticed that this problem is getting worse. I mean, it crashes on less and less demanding game. At the beginning, in january, it crashed while playing dishonored 2, but now, it crashes on rocket league...

***Well , it is not really at a random moment, usually it happens when I'm leaving a menu or a loading screen. But it can still be after 20 minutes into the game like it could be after 2 hours. Maybe this information can help.

I've tested my cpu, no problem.
I've tested my ram, no problem.
I'm constantly looking at the temperature of my cpu, my gpu and my mobo while playing games but there's nothing out of the ordinary.

Here are my components:
cpu: intel core i7-6700k 4.0Hz quad core processor
gpu: Evga GTX 1070
ram: crucial 2 x 8Gb
psu: OCZ ZX 850W 80+Gold certified Semi-modular
cpu cooler: cooler master hyper 212 evo
Using 2 monitors: 27'' 2k 75fps (my main), 21'' 1080p

I've built this pc in august so every part is less than a year old except for the psu that is almost 5 years old.

For the moment I don't play any game because I'm too scared it's gonna crash again.

Thank you for your time.

I've also opened it up and nothing seems to be out of normal. So I'm very confused right now. What is failing in this rig?

 

[Carnaxus]

Set up some kind of recording thing to watch the temperature monitor, preferably an external camera (so the footage doesn't get lost on crash), and wait for it to crash. Check to see if it records any kind of spike or drop, either in temperature or in operating voltage, just prior to the crash.

Also, check your Windows event log. There should be some reference to the crash in there.

One last thing. You said that you built this computer new in August, except for the PSU being almost 5 years old. There's a pretty good chance this is your PSU failing due to age.

 

[Jloch98]

Set up some kind of recording thing to watch the temperature monitor, preferably an external camera (so the footage doesn't get lost on crash), and wait for it to crash. Check to see if it records any kind of spike or drop, either in temperature or in operating voltage, just prior to the crash.

Also, check your Windows event log. There should be some reference to the crash in there.

One last thing. You said that you built this computer new in August, except for the PSU being almost 5 years old. There's a pretty good chance this is your PSU failing due to age.

I agree with Carnaxus. Another thing that might help, is to download a program called NZXT Cam. It monitors all the primary aspects of your computer, such as the CPU, GPU, RAM, and more. Might help to understand what part of your computer is running the hardest before the crash.

 

[Kuhulnim]

I'm currently using open hardware monitor to check the temperature of my cpu, gpu and mobo but i can't see what happened after the crash because it doesn't record anything. Does NZXT Cam record the state of the components in order to be able to see what happened before the crash?
I've also looked at the event log and the error that i receive at the moment of the crash is a critical event from kernel power. In my situation, is that supposed to help?

 

[Jloch98]

I'm currently using open hardware monitor to check the temperature of my cpu, gpu and mobo but i can't see what happened after the crash because it doesn't record anything. Does NZXT Cam record the state of the components in order to be able to see what happened before the crash?
I've also looked at the event log and the error that i receive at the moment of the crash is a critical event from kernel power. In my situation, is that supposed to help?

I'm not entirely sure if Cam can see what happened before the crash. I think all it really does it monitor the components. I would copy the error you get from the event log, and see if you can find something to fix it in Google. Sometimes a simple Google search can solve everything. Sorry I can't be of more help.

 

[Carnaxus]

The "critical event kernel power" error message is referring to you performing a hard shutdown; unfortunately, it's not the error that's causing the crash.

I have a theory, but I'm gonna need a bit more info about your computer first. What's your motherboard model? What speed is your RAM (DDR3-1866, DDR4-2400, etc.)? Exactly which EVGA GTX 1070 do you have (this can be found on the box the card came in)?

 

[Kuhulnim]

I looked for the error message on Google but I didn't find anything to solve my problem.

My GPU: EVGA GeForce GTX 1070 Superclocked 8Gb DDR5
My RAM: Crucial 16Gb Kit DDR4-2133 CL16 Sr X
My motherboard: gigabyte z170xp-sli DDR4 (lga 1151)

I've already tested my RAM by running a diagnostic to check my system for memory problem and it didn't give me any problem.
Furthermore, when I opened up my pc to check my PSU for something visually abnormal. I've noticed that on its fan there was a sheet with holes (like a net) (I wanted to post a picture of it but i don't know how). That net add some dust on it so i decided to wiped it out. While I was doing that, I wondered why this net is there so i just took it off. (This might be confusing but it's hard to explain without a picture 😛... if you really want a picture, just tell me how to post a picture that comes from my hard drive ). After that I decided to play some Outlast 2 and I didn't get any shutdown during that time. Could it be because removing that net helped a lot or because that game is not really demanding (for maximum settings, the game only needs a gtx 650)?

 

[Carnaxus]

The net could be a dust filter, and it could have been too clogged to allow proper airflow. Just how much dust was on it?

I was thinking maybe your PSU was either underpowered or dying, but an 850W PSU shouldn't have any issues with a computer that only pulls 362W. Try a heavier game with the net removed and see if it dies again.

 

[Kuhulnim]

The dust was covering all the part of the net and a little bit of the holes of the net (maybe 20% of the holes). I removed the net completely.
So I just tried a heavier game (Dying Light) where my gpu is used at almost 100%. I played for over an hour and didn't get any problem at all. At the same time, I was checking the temperature of my rig:
GPU: stable at 70 degrees celcius
CPU: not going further than 59 degrees
Motherboard: one temp is stable at 43 degrees and the other temp had a limit of 53 degrees
Note: for these temperatures, there is absolutely no exception, no weird peaks or drops.

 

[Carnaxus]

Yep, sounds like it was a dust filter, and it was clogged. I'd put it back on, and just remember to clean it periodically.

 

[Kuhulnim]

Sorry it took long before this reply. I just wanted to be sure that everything worked perfectly before picking the solution. After spending a lot of time on my computer, I didn't get any other crash so the problem really was the dust filter that had to be cleaned up.
Thanks for your great help guys!