Computer Freezes Randomly

[samcbshep21]

This is about my personal rig at home, but would appreciate any help.

 

So I upgraded my PC recently, changing the mobo, ram, cpu and cpu cooler. Other than this, the rest remained the same. I had no issues before my upgrade. Now seemingly randomly, my PC will freeze completely. Sometimes it will freeze and the sound will freeze with it leaving a loud buzz until I power the systemhttp://images.intellitxt.com/ast/adTypes/icon1.png off. I know its frozen as no keys respond to any commands (I tell by pressing the num lock key). The PC still has power to it as my LEDs inside, my Keyboard LEDs and fans all remain on. The only thing I can do is hold the power button down and then turn it back on.

 

This will happen from once a night to 3-4 times an hour. The only thing I notice is it only happens when gaminghttp://images.intellitxt.com/ast/adTypes/icon1.png and more specifically harder to run games. An example would be it happens when playing Minecraft with mods and shaders, but remove the shaders and its fine. Currently playing the division on high settings at 1440p and getting a solid 60fps all the time (temp readings below are from when playing that game). The obvious answer is a heat issue, I recently ran a second monitor so I can see my temps when the freeze occurs. around 45c (max 60c) on the cpu and around 65c (max 68c) on the gpu.

 

Specs:

-Win10

-Asus rog maximus viii

-Skylake 6700k (OC to 4ghz, issues happens when OC and not OC)

-cooler master nepton 240m

-16gb ddr4 ram

-corsair ax760i power supply

-gtx 980ti

-2 ssds, 1 hdd

 

Would really appreciate any help with this.

-

Thanks

[Norphy]

Those temperatures are fine, I've seen computers run much hotter than that and still be perfectly stable.

 

This is going to be one of those problems I'm afraid, something that's extremely difficult to pin down and resolve. The most obvious thing that I can think of is that it's either a RAM or a PSU issue.

 

How many sticks of RAM do you have in there? Does it still crash with 8GB or 4GB RAM? Do you have a spare PSU which you can test with? If not, is it any more reliable if you take an SSD or hard drive out or if you underclock the CPU?

 

I take it you've done the obvious, updated drivers etc?

[samcbshep21]

I have to sticks of RAM, 8gb each. I have moved them around with no success. I guess I should try it with just one stick but that'll take time to rule out each stick in each RAM slot.

 

I can't imagine it being the PSU unless it packed up just as I made the upgrades as there were no issues with my previous mobo and CPU. The only spare one I have is a 500w which might not be enough to run the rig at load.

 

Yeah all drivers are updated.

[Norphy]

I have to sticks of RAM, 8gb each. I have moved them around with no success. I guess I should try it with just one stick but that'll take time to rule out each stick in each RAM slot.

 

I can't imagine it being the PSU unless it packed up just as I made the upgrades as there were no issues with my previous mobo and CPU. The only spare one I have is a 500w which might not be enough to run the rig at load.

 

Yeah all drivers are updated.

 

Time is what it's going to take to sort this one out unfortunately. There isn't going to be a quick fix.

 

PSU is still an option, even if it did work OK before your upgrade. It could be that it's getting worn out or that it's just on the cusp of what it's capable of delivering. Just because it says 700W on the label doesn't mean that it's actually capable of providing that.

 

It may be worth using the using the paperclip trick to power all of your drives with the 500W PSU and running the motherboard and graphics card from the other one. If it's stable like that, you'll know it's a power issue.

[samcbshep21]

Somehow I knew that unfortunately.

 

I will try the RAM first as I will need to visit someone to get my PSU back to test and then will try your recommendation.

[ICTDirect_Dave]

This sounds like a PSU/GFX card issue. If it only freezes during gameplay when there's more load on the system then that would make most sense.

 

If it was a RAM issue you would expect it to crop up when you're using the machine for general browsing / video watching etc.

 

My gut feeling is that you might have a bad graphics card or at least a bad memory chip on the card. Have you tried removing the card and using the on chip Intel gfx to see if the fault persists?

[samcbshep21]

I have not tried that, how would I test the intel gfx to a similar level as it wont run anything near what the card will.

[JoeBloggs]

my first thoughts are PSU... I wouldn't use it until you've tested another high spec PSU. You could be damaging your kit if it is.

[Duranis]

Having exactly the same issue with my home rig at the moment (recently added SSD, new GPU and CPU, also have the same PSU I think), only seems to be on one specific game for me but I haven't really tried any others (benchmarking/stress test programs have run fine for over an hour though).

 

For this kind of fault it is normally temp, power, faulty hardware, in that order. Temps look fine on yours so next thing I would check is voltages/amps on the PSU. It might have been fine on the old rig but maybe the new CPU has a higher current draw which is enough to push the PSU into overload. You can try using something like HWmonitor to check voltages but seeing as it completely locks the PC when it happens it may not show you anything.

 

I would try disconnecting everything that isn't needed and has a power draw (including USB devices) and then underclock your processor and test the stability. If that seems to resolve the problem set the CPU back to stock and test again. If that is still ok reconnect everything else one bit at a time and test. If you get to a point that it becomes unstable again it may be that the last thing you connected has a problem, however it might also be that was just the thing that pushed the PSU into overload mode.

 

One other thing I would try doing is running some stress test programs to max out the CPU/GPU, disk and memory read/writes and see if you can force it to happen. If you can consistently make it happen on demand it makes trouble shooting a little easier.

 

Really frustrating problem to deal with as it is so inconsistent, thought I had solved it at the weekend as it run all day with no problems, then Monday night it crashed twice in an hour. Because it is just hard locking the system there is no good way (that I have found) to log exactly what is happening when it goes wrong.

 

I suspect that the problem for both of us is that the PSU needs upgrading to something with better ampage. It is probably working right at the limits when on full load and it just takes that one thing to push it over, hence why it is so inconsistent when it fails.

[Pyroman]

@samcbshep21 Could be related to this, what driver are you using? Have you tried rolling back the drivers or using the latest update they just released?

 

Updated: GeForce drivers 364.47 are causing catastrophic errors - PC Gamer

 

Nvidia GeForce 364.47 GPU driver is borking gamers' PCs - The Inquirer

 

edit: latest driver (beta) - http://www.nvidia.com/download/driverResults.aspx/99512/en-us

Edited March 11, 2016 by Pyroman

[Duranis]

My problem seems to have been my new graphics card (R9 380 so not the nvidia driver issue). Swapped it for my old card yesterday and it didn't crash once. New card is being sent back but because it was an exdisplay/open box one they probably wont have a replacement so will need to find another card for around £120-150 ish.

[difinity]

Was it sagging in the slot ? I had a PC in for repair recently that had a big heavy gfx card, pulled down in the corner. I used a cable tie to hold it up and it's not been back since.

[samcbshep21]

I have purchased a new PSU to see if that fixes the issue. Put it in last night and had no issues. Will let you know how it is after the weekend when I am more likely to be using the system for longer.

[samcbshep21]

Quick update - I still had the same issues with a new PSU. I decided to check my drivers (thanks to @Pyroman). I removed current driver using ddu in safemode. I installed a Nvidia driver going back to 358.50. With this driver insgtalled I have had no sytem freezes or crashes of any kind since. I am now going to put in my old PSU and if it works with that then I know that it was a driver issue and can return the new PSU.

[Pyroman]

Quick update - I still had the same issues with a new PSU. I decided to check my drivers (thanks to @Pyroman). I removed current driver using ddu in safemode. I installed a Nvidia driver going back to 358.50. With this driver insgtalled I have had no sytem freezes or crashes of any kind since. I am now going to put in my old PSU and if it works with that then I know that it was a driver issue and can return the new PSU.

 

Only just seen this, with the old PSU in are you all good now?

[samcbshep21]

Only just seen this, with the old PSU in are you all good now?

 

Sorry for the slow reply, I have been all good now since rolling back my graphics drivers and am using the old PSU.

[cpjitservices]

Quick update - I still had the same issues with a new PSU. I decided to check my drivers (thanks to @Pyroman). I removed current driver using ddu in safemode. I installed a Nvidia driver going back to 358.50. With this driver insgtalled I have had no sytem freezes or crashes of any kind since. I am now going to put in my old PSU and if it works with that then I know that it was a driver issue and can return the new PSU.

 

Why not keep the new one anyway PSU's can just 'die'... always good to have a spare one that will run your rig.

[samcbshep21]

Why not keep the new one anyway PSU's can just 'die'... always good to have a spare one that will run your rig.

 

You are completely right, it is just that I don't really want to be keeping spare parts lying around and because of Amazon Prime I can get a new PSU the next day. So as I don't need it right now, I would rather have my money back and it is less stuff lying around.

 

If it wasn't for Prime then I probably would have kept it as a spare.