PC restarts when playing a game

C

cardboardbox

Weaksauce
Joined
Apr 4, 2014
Messages
105
Hey guys,

I'm having a problem with my gaming PC, it restarts while playing games. :(

I was so sure the problem was the PSU that I got a replacement under warranty. At the same time I replaced my GTX 970 with a GTX 1060. Power consumption went way down and some games dont make my pc restart anymore (Fallout 4), while others do (Quantum Break). Again I figured the PSU was bad, so replaced it with a backup I had on the shelf. Same problem...

I did some stress testing (Prime 95, Aida64) without issue although I did not let them run for very long, maybe just 15 minutes because they made temps go much higher than games do. Generally a game causes a restart after 30-60 min. Neither CPU nor GPU fans are very loud during gaming, temps seem fine.

Could my motherboard be bad? I have a very large and heavy heatsink and because the ITX board only has support on the 4 corners, I think the middle of the motherboard is bowed downward slightly. I'll never get an ASRock mobo again, it was very flimsy feeling, but its what I have for now. Perhaps I should replace the HSF with the stock Intel HSF and see what happens? I think I have a Cooler Master Hyper T4 on my other PC in study. I could perhaps swap HSF since the Cooler Master is much ligher than the Deep Cool but nothing is going to be as light as the stock HSF I would assume.

i7-4790
ASRock H97M ITX
PSU: EVGA 600 B1 or Corsair 430W
G.Skill Ares DDR3 2400 16GB
MSI GTX 1060 Gaming
Deep Cool mc6002gs HSF: http://www.xoxide.com/logisys-deepcool-gamerstorm-cpucooler.html
Fractal Node 304 ITX case
 
Try a stress test with Furmark. Don't swap back to the stock HSF, they're terrible especially in an ITX environment.
 
sinisterDei said:
Try a stress test with Furmark. Don't swap back to the stock HSF, they're terrible especially in an ITX environment.
Click to expand...
CPU burn was fine for 30 min but the GPU test caused my pc to shut down after about 16 min. At 14 min I checked and it was running fine and the temp was 74 which is much hotter than it ever gets during normal use. With the Corsair PSU my PC shuts down during gaming/Furmark. The EVGA PSU would shutdown but then come back on and so I called that a restart although maybe it wasnt really.

Since I was having this problem with the previous GTX 970 and now its happening with the 1060, you'd have to think its not the video card, right?
 
Well, my #1 go-to would be the PSU, except you've tried two.

As for the GPU, 74 C is not unreasonable for a GPU under furmark - at that point the card is likely throttling itself a bit. But it shouldn't be unstable at that temperature.

Considering it's stable in normal 2D operation, I'm tempted to blame the motherboard, but I certainly can't prove it.

Try a test- play an older game, something you would easily get 150+ FPS in like Half Life 2 or something, but enable vSync (presuming you have a 60hz monitor here) which caps the framerate at 60fps. This will have the effect of using your GPU in 3D mode, but not in a stressful manner, it might be running at 20% load. I'm curious if it locks up / reboots in those circumstances.

As to ASRock mobos, I have four of them in my house. They've always been stable, though I've never mounted anything heavier than a CM Hyper 212 on them.
 
Just to ask a follow-up question, I'm presuming you've reinstalled Windows at some point, and that's why you're looking at the hardware?
 
maybe your gpu needs a little tlc. pull it out and take off the heat-sink, give it a good cleaning and new paste, check it over for any signs of overheating or damage. make sure the power sockets are clean and the pins are good.
 
sinisterDei said:
Well, my #1 go-to would be the PSU, except you've tried two.

As for the GPU, 74 C is not unreasonable for a GPU under furmark - at that point the card is likely throttling itself a bit. But it shouldn't be unstable at that temperature.

Considering it's stable in normal 2D operation, I'm tempted to blame the motherboard, but I certainly can't prove it.

Try a test- play an older game, something you would easily get 150+ FPS in like Half Life 2 or something, but enable vSync (presuming you have a 60hz monitor here) which caps the framerate at 60fps. This will have the effect of using your GPU in 3D mode, but not in a stressful manner, it might be running at 20% load. I'm curious if it locks up / reboots in those circumstances.

As to ASRock mobos, I have four of them in my house. They've always been stable, though I've never mounted anything heavier than a CM Hyper 212 on them.
Click to expand...
A CM Hyper 212 is less than half as heavy as my HSF. Mine needs to lose some weight but it didnt make my previous mobo bend in the middle because the board was stronger (thicker?).

I'll try your Half Life 2 idea a bit later if this thing dies on me again. Right now I'm an hour into a Furmark stress test. Max temp is 70. The only thing different this time is that I took the box out of the living room tv stand where it is mostly enclosed on three sides except for a slot in the back for cables. Its now in the study out in the open on a desk. I'm using DVI instead of HDMI and a different mouse/keyboard/power cable. I know, really stretching there looking for answers. :)

sinisterDei said:
Just to ask a follow-up question, I'm presuming you've reinstalled Windows at some point, and that's why you're looking at the hardware?
Click to expand...
I haven't done any Windows reinstall since I did the Windows 10 upgrade in July I think. Hoping to avoid a reinstall...

pendragon1 said:
maybe your gpu needs a little tlc. pull it out and take off the heat-sink, give it a good cleaning and new paste, check it over for any signs of overheating or damage. make sure the power sockets are clean and the pins are good.
Click to expand...
I'll try that when I'm a bit more desperate since I've never removed a heatsink from a gpu. :)
 
If you've got a spare hard drive laying around, and you want to see if the Windows install is the problem, you can always temporarily disconnect your live OS and data drives and plug in the temporary one and install on that to test stability.

That said, if the system is stable in the study- which I presume is a different physical location in the house- then I would also presume it's plugged into a different electric socket. It could be a problem at the wall, rather than at the PSU.
 
The restart to desktop sounds like it's a software, OS, or driver error. Have you tried to do a CCLeaner type thing to your registry? System restore? I'm noobish compared with the elders here but the simple stuff often clears a lot of things. Goodluck and I hope for your resolution.
 
cardboardbox said:
A CM Hyper 212 is less than half as heavy as my HSF. Mine needs to lose some weight but it didnt make my previous mobo bend in the middle because the board was stronger (thicker?).

I'll try your Half Life 2 idea a bit later if this thing dies on me again. Right now I'm an hour into a Furmark stress test. Max temp is 70. The only thing different this time is that I took the box out of the living room tv stand where it is mostly enclosed on three sides except for a slot in the back for cables. Its now in the study out in the open on a desk. I'm using DVI instead of HDMI and a different mouse/keyboard/power cable. I know, really stretching there looking for answers. :)

I haven't done any Windows reinstall since I did the Windows 10 upgrade in July I think. Hoping to avoid a reinstall...

I'll try that when I'm a bit more desperate since I've never removed a heatsink from a gpu. :)
Click to expand...
I just re-read the thread and scratch what I said about cleaning. its a new card. durr...
so you've got a new card, new gpu and the system still looses power under heavy gaming? maybe disable any warning/shutdown temps in the bios. it could be board flex, you could try your stock hsf if you haven't yet. it kinda seems like a short somewhere too. if you not opposed to it maybe pull it all out of the case, check it over and test it outside of the case. I know its a pita but so is a non working system!
 
sinisterDei said:
If you've got a spare hard drive laying around, and you want to see if the Windows install is the problem, you can always temporarily disconnect your live OS and data drives and plug in the temporary one and install on that to test stability.

That said, if the system is stable in the study- which I presume is a different physical location in the house- then I would also presume it's plugged into a different electric socket. It could be a problem at the wall, rather than at the PSU.
Click to expand...
I was thinking along the same lines and just swapped the power cord. Which makes no sense to me but I just made it through two stress tests of 30 min and 60 min. Max temp on the 60 min was 75. Yes, study is a different location. Gaming rig is in the living room.

Nanogrip said:
The restart to desktop sounds like it's a software, OS, or driver error. Have you tried to do a CCLeaner type thing to your registry? System restore? I'm noobish compared with the elders here but the simple stuff often clears a lot of things. Goodluck and I hope for your resolution.
Click to expand...
It actually shuts down and then turns back on. Doesnt ever get back to desktop until Windows loads again. System restore is on my list of ideas that I'm working my way down to.

pendragon1 said:
I just re-read the thread and scratch what I said about cleaning. its a new card. durr...
so you've got a new card, new gpu and the system still looses power under heavy gaming? maybe disable any warning/shutdown temps in the bios. it could be board flex, you could try your stock hsf if you haven't yet. it kinda seems like a short somewhere too. if you not opposed to it maybe pull it all out of the case, check it over and test it outside of the case. I know its a pita but so is a non working system!
Click to expand...
Yeah thats probably the last thing I'll try because its such a pain in the ass but I cant find my mobo serial number for warranty claim and I need to check if there's a sticker on the bottom of it.
 
Run multiple instances of HCI Memtest, which is generally a better memtest than Memtest86 or Memtest86+ (though no software RAM tester can be perfect). Since you have 16GB of RAM, you should run 7 instances of 2047 MB each of HCI Memtest. If the RAM is bad, errors should start showing up in fairly short order, though it needs to be run for many hours before being considered golden (have some patience, do it longer than 15 minutes is what I'm saying).

If you get errors with any memtest, make sure the motherboard's BIOS is updated to the latest version and retest after updating if it wasn't. Many BIOS updates are to increase compatibility with RAM.
 
The system might simply be BSOD and it's auto-restarting so you don't see it. Go to system properties and check it out- I've included the location in the attached image. Uncheck the box and if it BSOD, it'll just hang there rather than restart.

But since you made it nearly 90 minutes combined of testing in the other room, power is back on the menu boys.
 

Attachments

  • Auto_Restart.PNG
    Auto_Restart.PNG
    33.8 KB · Views: 45
Have you messed with any settings in the bios for overclocking? I messed with mine at one point and forgot I did and when my pc began running like crap I thought things were wrong with hardware. I eventually loaded default settings in my bios and it worked out. My board is a Asrock running for 5 years now.
 
Ever since I swapped out the power cord I haven't had a problem. I'll wait a few more days before I'm convinced its fixed. What a strange fix, dont think I've ever had this happen before.
 
Once, about two years ago, in the middle of using my PC, I started hearing this weird hissing noise and then a pop. Nothing had happened - my PC was still on, nothing had anything changes at all, I just heard the noises.

I got to looking around, I found a standard black power cord like you use for your PC or monitor. At some point in it's life, I had unplugged it from any equipment, but left it physically plugged into the power strip. No big deal right?

Well, this day, I guess it decided "fuck this shit I'm out of here" because it seemed to have physically melted internally, and then externally a bit. The length of the entire core was stiff and brittle, no longer flexible, and any attempt to bend it resulted in bits of the cord literally snapping apart. And there was a spot against the wall, I guess the epicenter of whatever bad thing had happened, where there was a scorch mark against the wall.

The moral of the story is, power cords are prone to occasional betrayals and will try and fuck your shit up, so don't trust them!
 
sinisterDei said:
Once, about two years ago, in the middle of using my PC, I started hearing this weird hissing noise and then a pop. Nothing had happened - my PC was still on, nothing had anything changes at all, I just heard the noises.

I got to looking around, I found a standard black power cord like you use for your PC or monitor. At some point in it's life, I had unplugged it from any equipment, but left it physically plugged into the power strip. No big deal right?

Well, this day, I guess it decided "fuck this shit I'm out of here" because it seemed to have physically melted internally, and then externally a bit. The length of the entire core was stiff and brittle, no longer flexible, and any attempt to bend it resulted in bits of the cord literally snapping apart. And there was a spot against the wall, I guess the epicenter of whatever bad thing had happened, where there was a scorch mark against the wall.

The moral of the story is, power cords are prone to occasional betrayals and will try and fuck your shit up, so don't trust them!
Click to expand...
You should have told him this in the first place!!! Duh!! LOL
 
sinisterDei said:
The moral of the story is, power cords are prone to occasional betrayals and will try and fuck your shit up, so don't trust them!
Click to expand...
no more restarts, so I'm pretty certain it was the power cord. It wasnt visibly damaged like yours was and still works fine for light duty. Funny thing is, if EVGA had told me to try a different power cord before approving the RMA (and they paid shipping both ways), I would have been annoyed with them for telling me to do something so stupid. Learn something new everyday I guess...

thanks for your help, end everyone else also!
 
Glad to hear it, and doubly glad it was a $2 solution and not a $200 solution.

And yes, I would also have rolled my eyes at a company if they asked me if I'd tried to use a different power cord. I probably would just have said Yes whether I had or not.
 
IYO...
ive seen it, not often but ive seen it. even 2 prong cords for laptops, brick is fine cord is dead.
 
I've also seen power cords flake out. Most common is when the plug on the cord that plugs into the power supply itself is no longer a snug fit.

If you have ever actually cut the ends on a power cord apart, it would surprise you how cheaply they are made. I have no idea how they actually pass UL certification.
 
I read a story once that a lot of them don't, and some manufacturers just stamps the UL logo on there and hope nobody checks. That could be bullshit though.
 
well a power cord IMO would mostly be easy to detect....in other words it would turn off if the cord was bumped....So thats why i figure its something else. OR the cord would be pretty hot in the bad spot on the cord....
 
Well, a fix is a fix. Here in my work, we sometimes get machines that would intermittently work, and for medical equipments, that is unacceptable. We have bad power cords now and then, due to nurses and medical practitioners, asshole patients and their family, that tug the wire out of the outlet instead of pulling from the plug.
 
I would guess that there was a partial short in it and when under heavy load the current draw was enough to cause an arc and a reboot.
 
primetime said:
I seriously doubt it was the power cord....IMO
Click to expand...
I wouldnt believe me either. But I'd tried the obvious stuff and at that point I was out of good ideas so I started with the easiest dumb idea I had. Maybe not completely a dumb idea since the problem I thought was caused by the PSU so this was pretty close to it.

primetime said:
well a power cord IMO would mostly be easy to detect....in other words it would turn off if the cord was bumped....So thats why i figure its something else. OR the cord would be pretty hot in the bad spot on the cord....
Click to expand...
I didnt think to check what the cord felt like. I didnt really expect this to fix it.

pendragon1 said:
I would guess that there was a partial short in it and when under heavy load the current draw was enough to cause an arc and a reboot.
Click to expand...
that sounds most likely. With my GTX 970 it was rebooting in Fallout and Quantum Break but after I swapped the 970 for a GTX 1060, Fallout 4 stopped causing reboots. Using a watt meter, I found that QB used a lot more watts than FO4 and the switch to a more efficient GPU must have lowered FO4's power consumption under the threshold that caused the problem.
 
primetime said:
I seriously doubt it was the power cord....IMO
Click to expand...
you're right, my pc is restarting again. I dont understand why it took 2 weeks to happen again.

Right now I'm thinking about moving everything to a different case thats easier to work in. My ITX box is so cramped, annoying to work on it. I have an Antec P50 in my closet that I could use. Maybe just moving everything will help me figure out the problem.

One stupid idea I have is to remove the bluetooth wires/antennas. My motherboard has terrible antennas that dont screw in well and I think I've spun them around so many times that I might be twisting the connection on the mobo too much. Yeah, told you it was a stupid idea. :)
 
Step One...Make a Sig with all hardware being used.....Otherwise we have No idea what your using or how or how to give advice! Take pictures and post them....Hey If your doing something stupid you need to know! Also post you bios settings and OS being used.....Stay calm cause we figure this out for a living lol;)
 
Which version of Windows?

You can view the event log in Windows at the time the restart happens and see why it rebooted. I know in Win7 their is a stupid setting that will tell Windows to automatically restart in the event of a system failure and that doesn't always mean it needs to be restarted. I imagine Win8 or 10 has a similar option as well.

Either way, the Windows Event log should shed some light on why it's rebooting.
 
I had almost the exact same thing happen to me.

After replacing everything - and I mean everything, I never did figure it out. I thought I had it narrowed down to the CPU, except that I'm still using it in a build and it works fine now.

I have an entire second computer now of all the parts. I am still using all the original parts, minus 1. Both systems seem to work fine. I did a lot of swapping back and forth to try to duplicate it, and while at various times one or the other system would randomly restart. The only thing I was able to get halfway consistently was whichever system had a Corsair H110 installed as the heat sink would tend to, but not always, be the system that would have the reboot issue. So I replaced that with a CM212.

I don't know how or why, but it works now.
 
thanks for the ideas, I will report back when I have more info and have tried everything suggested. Will also make a sig soon. :) Busy today with the kids and getting my car fixed today and then visiting the grandparents.
 
One time i had a restart issue that seemed to happen when i bumped the case....It turned out to be the sound card was not fully seated...It was pretty easy to figure out once the sound stopped working.
 
ok, updated my sig.

I tried switching my components to a new Antec P50 case I had in the closet. I discovered that my heatsink was about 3-4 inches too tall but at that point I thought I'd leave the case door off since I was just testing. But then the video card also wouldnt fit so the P50 was tossed back in the closet.

Then I tried an ancient case I had laying around, I think its a small full ATX mobo sized case. That one also didnt work because it was an inch too small for my video card, though the case is pretty big overall.

So back to the original ITX case but this time I swapped the HSF with the stock intel cooler. So far some stress testing was successful and no problem playing a game but this time I'm going to wait at least 2 weeks before I start thinking it was a success. I tried to put an old Zalman HSF on but the backplate would not fit, possibly because of this: http://imgur.com/a/wX3qF

Does it seem curved to anyone else? Perhaps the angle I took the picture at but to my eye it looks curved. I wonder if I warped the board with the previous heavy HSF that I maybe put on too tight?

btw - I hate the whine from the stock cooler!
 
cardboardbox said:
...possibly because of this: http://imgur.com/a/wX3qF

Does it seem curved to anyone else? Perhaps the angle I took the picture at but to my eye it looks curved. I wonder if I warped the board with the previous heavy HSF that I maybe put on too tight?
Click to expand...
Looks bent to me too. Judging from the shadow between the plate and the board being equidistant, I'd guess that the board might be bent too.
 
update:

While gaming, I've had no restarts but yesterday I did a stress test with Furmark, left the room for 30 min and when I came back my pc was off so there's still an issue. Judging by fan noise, I'm probably playing less stressful games lately and that might be why I haven't had restarts/shutdowns for the past 2+ weeks while gaming.

Today when I turned on my pc I had a red X on the network icon in the system tray, it was like that component just disappeared, no network adapter was showing up in device manager. I tried reinstall drivers but that didnt work. Even though I'm on the latest BIOS, I flashed it again and upon restart my network adapter was back.

I've read the stories about how bad ASRock is for RMA'ing so I think I'll start looking around for a good deal on a new motherboard. I'm worried that one of these restart/shutdowns is going to break one of my other components. I think ASRock will look for any reason to deny an RMA claim and since I think my motherboard is slightly warped, I'd rather not throw away $12+ on shipping and nothing get fixed.
 
is it warped around the cpu socket, maybe you hsf is too tight? if so, loosen it a bit and see if problems go away. seen others complain aboot this with asrock boards...
 
You must log in or register to reply here.
Back
Top