Weird PC Restart Problem

[marmiteonpizza]

Okay this is going to take a while to explain, but I need all of your help.

Basically, yesterday in the middle of a Civ V game, my computer randomly and for no apparent reason restarted (resulting in me losing at least 2 hours of saved gameplay unfortunately).
The same thing kept happening with a lot of different games which were more graphically intense than Civ, so I immediately assumed it was to do with my graphics card. Plus, I had just updated my nVidia driver so I was certain it was my graphics card.
I uninstalled all of my current and previous graphics drivers, got the disc which came with my graphics card and installed the driver from there again. My PC still restarted when I played games on the original driver, even though I had used that driver for weeks with no problem.
So after it restarted again, I installed the latest driver from the web, played Dishonored for about an hour and thought the issue was solved. Guess what?...It restarted. Only this time, the monitor wouldn't pick up any signal whatsoever from my PC and for some reason, the power/restart button was completely unresponsive. Even if I held the power button down for ten seconds, it wouldn't power off.

So I turned it off at the socket, checked my HDMI connections - all were fine.
I decided I would take a look inside my PC while I turned it on to see if any of the status LED's on my motherboard were red. First thing I noticed was that the processor heatsink fan was not working, and sure enough, the LED for the CPU on the motherboard was red.
So I checked all the power cables going into the motherboard and power supply for the processor - they seemed good. Then I only unhooked the catches for the heatsink fan but didn't take it out, rehooked them and started the PC up again. Red LED light came on for a second, disappeared and the heatsink fan was working.
My PC came up on the screen as normal and is working.

I haven't yet tried to play a game since, and will do so in not long, hoping it won't suddenly restart at some point.
Obviously the fault is/was with the processor, does anyone know what might be/have been wrong?

 

[Recca29]

can be heating problems. can u check the temps of the processor and graphics card while running some benchmarks.

 

[marmiteonpizza]

can be heating problems. can u check the temps of the processor and graphics card while running some benchmarks.

How do I go about this for the processor?

 

[RCoon]

Obviously the fault is/was with the processor, does anyone know what might be/have been wrong?

Fault looked like the CPU fan was not spinning, therefore the CPU was slowly but surely heating up until it reached the TJ MAX and shut down to prevent it from frying. I assume you've tried playing games with the side of the case off to check that the CPU heatsink fan is actually working? Might be a place to start. Also try the CPU Fan 3pin in the CPU_FAN or CHASSIS_FAN pin headers on your motherboard. Could be a bad header.

If it is spinning, you'll want to run a temperature monitor. HWInfo does this, as does SpeedFan and RealTemp. Use any of them. Your processor should not go above 63 degrees. I believe 65 is max, but 63 has always been a safe point for me on AMD FX processors. If it's hitting 63 degrees at stock, it's time to repaste the CPU and reseat the heatsink, perhaps even buy a new heatsink (Hyper 212 EVO is a very good budget cooler for stock/minor OC).

 

[marmiteonpizza]

If it's hitting 63 degrees at stock, it's time to repaste the CPU and reseat the heatsink, perhaps even buy a new heatsink (Hyper 212 EVO is a very good budget cooler for stock/minor OC).

Surely I mustn't need to replace my paste already?! It's only been two to three weeks since I built this PC!
But I haven't taken the case off while playing a game, only while booting up. So I'll try that, play Dishonored for a while and see if the fan stops at all =) And I may as well use one of those temperature monitors at the same time.

Got to hoover my room first though Will update when I've tried this out.

 

[RCoon]

Surely I mustn't need to replace my paste already?!

I doubt it, but in the event the CPU Cooler is mismounted (check that it's nice and flat and flush), you'd probably need to repaste. More than likely the fan issue though.

 

[Recca29]

I use HWMonitor. its free. you can download it from http://www.cpuid.com/softwares/hwmonitor.html.

gives you temps for many components + there min max temps.

Edit:
and as @RCoon said >63c is bad. take his advice and change that stock cooler

 

[marmiteonpizza]

UPDATE

Put PC on its side with cover off and downloaded/ran one of the monitor software while playing a game for a good half an hour and the CPU fan was constantly on. Of course, the software read that the temprature was no way near the max. due to this.
So everything seems to be running as normal again, but I'll be playing games throughout the day and will update further if the problem falls upon me again.

Just a question: why did the PC only restart when I was in-game? Why not any other time? Do you think the fan was off the whole time (even when not in-game) and the CPU only overheated when it needed more power for the games?

 

[RCoon]

the CPU only overheated when it needed more power for the games?

This. Your CPU can typically run without the fan turned on, as it's not drawing much power when it has no load. Once a CPU is loaded, it will draw more and more power and put out more and more heat. If the fan is not removing the heat, then it just builds up until it eventually hits the max, and causes a shutoff.

 

[Aquinus]

Side note: Civ 5 has auto saving turned on by default. You may have only lost a few turns of your game. This has happened to me a few time and have always been saved by that.

Also you could be BSODing without knowing it. It may be setup to restart automagically. You can find records of BSODs in the event viewer or by disabling auto restart as shown below.

Additionally, using too much paste can also cause temps to run high.

 

[marmiteonpizza]

Side note: Civ 5 has auto saving turned on by default. You may have only lost a few turns of your game. This has happened to me a few time and have always been saved by that.

Unfortunately I've turned autosaving off due to hate having too many save files. So I've got to go back a pretty long way! Will try and get back to my previous place today.
Thanks for the idea, but it was nothing to do with BSOD, as the restarting was completely random and I saw for a fact that the CPU fan was not working.

I think it's all sorted out now, probably a loose connection on the motherboard that I fixed when I replugged cables this morning.

 

[marmiteonpizza]

You guys will not believe how this issue has escalated! I had an absolute nightmare today fixing it.
Basically, it was the stock CPU fan which was naff, so I had to order a Cooler Master one which was meant to arrive Saturday but didn't arrive until today. Then, I found out that it was too big to install with the motherboard still in the case, so I had to completely strip my PC! Motherboard out, all wires unplugged...everything! Just to fit this blessed fan inside.
Finally got it all sorted and put it all back together, so relieving!

 

[Recca29]

You guys will not believe how this issue has escalated! I had an absolute nightmare today fixing it.
Basically, it was the stock CPU fan which was naff, so I had to order a Cooler Master one which was meant to arrive Saturday but didn't arrive until today. Then, I found out that it was too big to install with the motherboard still in the case, so I had to completely strip my PC! Motherboard out, all wires unplugged...everything! Just to fit this blessed fan inside.
Finally got it all sorted and put it all back together, so relieving!

What are the temps after the upgrade.

 

[Jborg]

Sounds like the CPU was heating up and causing your computer to shut down possibly.

Should try to test temps with new cooler. I swapped to the Hyper 212 EVO on my 8350 and noticed an instant 25c drop in temps while under load.

If that fan isn't fitting easily tho, not sure if the HYPER 212 would, its pretty big. It comes about 8MM from touching my RAM.

 

[marmiteonpizza]

Sounds like the CPU was heating up and causing your computer to shut down possibly.

Should try to test temps with new cooler. I swapped to the Hyper 212 EVO on my 8350 and noticed an instant 25c drop in temps while under load.

If that fan isn't fitting easily tho, not sure if the HYPER 212 would, its pretty big. It comes about 8MM from touching my RAM.

Yeah I got the same fan, and it did fit but the reason I had to take my whole PC apart was because there was NO way I was going to fit the metal plates (that clamp the heatsink down) while all the cables were everywhere and my motherboard was in the case. So out came the motherboard, graphics card, wireless adapter and all the cables -.-

What are the temps after the upgrade.

Temp is only averaging at 12 degrees Celsius when idol now =) Compared to in the thirties with the stock heatsink!

 

[Jborg]

You should still test temps out under load....

That still is a significant drop in idle temps tho.

 

[FX-GMC]

Fault looked like the CPU fan was not spinning, therefore the CPU was slowly but surely heating up until it reached the TJ MAX and shut down to prevent it from frying. I assume you've tried playing games with the side of the case off to check that the CPU heatsink fan is actually working? Might be a place to start. Also try the CPU Fan 3pin in the CPU_FAN or CHASSIS_FAN pin headers on your motherboard. Could be a bad header.

If it is spinning, you'll want to run a temperature monitor. HWInfo does this, as does SpeedFan and RealTemp. Use any of them. Your processor should not go above 63 degrees. I believe 65 is max, but 63 has always been a safe point for me on AMD FX processors. If it's hitting 63 degrees at stock, it's time to repaste the CPU and reseat the heatsink, perhaps even buy a new heatsink (Hyper 212 EVO is a very good budget cooler for stock/minor OC).

70 Degrees is the end of the Thermal Margin in AMD Overdrive. If AMD says its ok I'm inclined to believe them. So 70 degrees (or 0 degrees thermal margin) is max. Trying to stay under 65 degrees (or 5 degrees thermal margin) is always a good idea like you say.

 

[marmiteonpizza]

You should still test temps out under load....

That still is a significant drop in idle temps tho.

How do I set up a test for when I'm in game or something? To actually record the results instead of just showing a live feed? I've got HWiNFO but can't find out how to run tests.

 

[Jborg]

Just monitor it by checking the temp every so often. If you reset the Min/Max in HWMonitor before you start the game that will provide a better reading. You dont really need to graph it, just make sure its running cool enough, and obviously not causing restarts anymore. Like FX-GMX said above, if your in the 60-70 temp range, that is near its max safe temps.

You may also want to get AMD Overdrive. this provides the Thermal Margin reading which is tells you how close the CPU is to its max safe operating temp. (Better for reading temps under load - from what I have read)

But honestly, I have the same 8350 as you, and while playing BF4 on max detail, even after 2-3 hours I am only at max temp of 45-52c package temps, then again I do have the 212EVO.

Somebody correct me if I am wrong, but I believe the max safe Socket Temp is like 70 or 71c, and the max safe CPU Package temp is like 63c I believe. There are 2 seperate readings in HWmonitor for these.

 

[OneMoar]

a hyper212 may not be enough for a 8350 depending on your ambient temps
I would definitely get a second blade master fan and run push pull
http://www.newegg.com/Product/Product.aspx?Item=N82E16835103069
its completely normal to need to remove the motherboard to install a cooler if your case doesn't have a cutout
the stock amd cooler exists simply to get the machine posting they are worthless for anything other then that unless you live in the arctic circle

 

[Jborg]

The Hyper 212 EVO works well enough for me. Then again all I do is game with some minor Sony Vegas and Photoshop usage.

My max temps running Prime @ 4.4Ghz was like 55c if I am remembering correctly (This was only for like an hour). I would say ambient temps in my house are around 65F, which is like around 25-30c(Edit: 21c ambient average) I believe. (Sorry for being a Farenheit nub)

But really, the HYPER 212 EVO gets the job done + more. At least for me.

 

[OneMoar]

I would be very surprised if a Evo managed to keep a 8350 under 60c @ full load
you aren't running the cpu hard enough to hit peak thermals

 

[Jborg]

Well it works great for me. Not one problem with being to hot. Unless your going to go crazy with OCing. I personally believe the 212 EVO is enough.

 

[Jborg]

I would be very surprised if a Evo managed to keep a 8350 under 60c @ full load
you aren't running the cpu hard enough to hit peak thermals

This is roughly 50-55 minutes into running Prime95 on Small FFT's with the 212 EVO @ 4.4Ghz, currently I have it at 4.3Ghz though. But it seems fine to me...? Other than that error.

The temps seem reasonable....

Not trying to jack the thread. Do hope you fix the issue.

 

[OneMoar]

This is roughly 50-55 minutes into running Prime95 on Small FFT's with the 212 EVO @ 4.4Ghz, currently I have it at 4.3Ghz though. But it seems fine to me...? Other than that error.

The temps seem reasonable....

I see a bunch of bunk sensor readings so who knows but its certainly possible tho if your ambient temps are low and your case has great airflow and you have a cooler running chip then most samples I have seen reviewe

also if you are getting errors under prime your oc is not stable ... you have some room with voltage but not temps I would be getting some better cooling or at least a second fan ..
the difference between the evo and non evo is 3 to 5C you are pushing toward 60 and then there is the fact that there is quiet a bit of fudge room going on with the TDP of the 8350