Computer freeze at games

[coolkingler1]

I may have put this in the wrong sub forum, can a mod move it if so?

I made this computer for my friend last year. Let me start with the parts:

Parts

CPU: AMD x4 II965 AM3 Black edition at 3,4 Ghz
Motherboard: Asrock M3N78D
PSU: Corsair CX500
GPU: GTX 550 TI
RAM: 2x4=8GB Corsair Vengeance
HDD: WD 500GB Caviar Blue

Story

Let me start at the beginning:

The pc was working fine at first, later on it showed symptoms and then it just started being very bad.

I went over once last year to check it out and it turns out his CPU fan was half off, so I put it back on. It was fine again and LoL ran smooth again.

A while ago he started complaining about stutter in League.
Also his stock fan makes a LOT of noise, but it does work as intended.


The stutter in League of Legends was unexplainable, his FPS was more than 60 at benchmark settings, the graphics weren't even too high for the sake of testing. The stutter was always there. It was not a big issue since he could play without issues at all, but it was annoying.


So today I went over with an old GTX 460. It did not work as it displayed an error message so we put the GTX 550 TI back in again, the 550 TI is not broken at all, we just thought we would test another video card. That GTX 460 is just dead.

We also took off the CPU cooler, wiped it clean from the old thermal compound it came with and did the same with the CPU, we put a bit on the CPU again of new thermal compound and spreaded it nice and thin. We put the CPU cooler back on.

I updated the video card driver to make sure everything was up to date.

Well it didn't work, too bad.

Next problem

Then I started to work on the sound problem.

I went to the mobo settings, I Enabled ' Quiet CPU mode' (or something of that nature, allowing me to set a target CPU temp and control the CPU fan speed.

The settings were at 50C target temp and level 9 fan speed. (the max)

I changed it to 57 target and 7 fan speed. (it did reduce the noise a lot)

Then I restarted with the new settings, everything worked. Then I started League of Legends. As soon as the loading is done and you get ingame you have 2-5 seconds before the entire computer freezes completely, no choice but to restart via the buttons.

I tested it with Skyrim, same thing, instant freeze.

Coretemp showed it froze at about 50C when I enabled CPU Quiet mode but did not change the target temp to 57 yet.


So I changed the settings back, though I am not quite sure I disabled ' CPU Quiet mode' again.

Now he can't play games.


What could be the problem here?

I am not sure if Target Temp means it will just freeze if it goes over that temp?

Once again I am not sure I disabled Quiet CPU mode.

 

[mindless1]

When a problem only surfaces in a specific game, that usually means the game is buggy and you have to wait on a patch... or, try more games to see if you can find any common denominator, though you might touch-test the heatsink (or use software to find) video card temperatures in case it is throttling back from excess heat.

No, target temp would not freeze a system. It only controls fan speed response to temp and that temp is fine. Even the other older method of setting system fault from CPU overheating would shut down the system, not freeze it.

How old is the PSU? If not under warranty I'd pop it open and check for vented capacitors.

 

[coolkingler1]

When a problem only surfaces in a specific game, that usually means the game is buggy and you have to wait on a patch... or, try more games to see if you can find any common denominator, though you might touch-test the heatsink (or use software to find) video card temperatures in case it is throttling back from excess heat.

No, target temp would not freeze a system. It only controls fan speed response to temp and that temp is fine. Even the other older method of setting system fault from CPU overheating would shut down the system, not freeze it.

How old is the PSU? If not under warranty I'd pop it open and check for vented capacitors.

It is under warranty, it is not old at all, it is almost a year old.

It is not LoL only, I tried Skyrim like I said in the OP but it happened there too.

I will look out for the video card heat, however, it was working fine before.

 

[mfenn]

Try testing with just one memory stick (cycle each through individually).

 

[coolkingler1]

Try testing with just one memory stick (cycle each through individually).

I will try this next time.


So far we have:

-Check GPU temps
-Check RAM sticks


What else can I add?

 

[mindless1]

Check PSU voltage and ripple change between idle and game-crash condition. Pull case side panel off and point a desk fan at it to see if that helps, though I think it is far more likely a power issue than heat, as it crashes so soon after beginning a game, unless a heatsink had entirely fallen off.

 

[coolkingler1]

Check PSU voltage and ripple change between idle and game-crash condition. Pull case side panel off and point a desk fan at it to see if that helps, though I think it is far more likely a power issue than heat, as it crashes so soon after beginning a game, unless a heatsink had entirely fallen off.

Okay thank you, how do I check voltages?

 

[mindless1]

multimeter

 

[mfenn]

Check PSU voltage and ripple change between idle and game-crash condition. Pull case side panel off and point a desk fan at it to see if that helps, though I think it is far more likely a power issue than heat, as it crashes so soon after beginning a game, unless a heatsink had entirely fallen off.

You're not going to be able to measure ripple unless you have a 'scope. Checking that the voltages are in range is a good idea though.

 

[mindless1]

^ Without a scope you won't get a simultaneous frequency reading but yes you can measure AC (hint) ripple voltage on a DC rail with a multimeter.

 

[mfenn]

^ Without a scope you won't get a simultaneous frequency reading but yes you can measure AC (hint) ripple voltage on a DC rail with a multimeter.

Measuring the amplitude of the peaks tells you pretty much nothing though. Especially not a time-averaged amplitude like you'll get on a DMM.

EDIT: OK, I guess it would tell you "holy shit somethings wrong" if the amplitude were really high. If things look OK on the multimeter, that doesn't necessarily mean that it's all good though.

 

[mindless1]

If the amplitude and DC voltage measurement are both within bounds, it tells you there's clean enough power getting to the load, the PSU is ruled out. The problem is without data acquisition/logging, you might not catch that moment that caused the freeze, but you might catch it acting badly before a freeze... and some meters will do logging or max/min value.

 

[coolkingler1]

I don't have that meter.

I will go and check tomorrow though.

He does say the GPU is getting very hot.

 

[coolkingler1]

I solved it, all I did was take out the CMOS battery and put it in again. Boom, games worked.

 

[Revolution 11]

It always surprises me how many problems are solved by simple solutions like this one.

 

[Torn Mind]

Very interesting case we have here.

 

[coolkingler1]

Yeah, it surprised me too. Guess it was of the 460gtx exchange.

 

[coolkingler1]

I knew it was too good to be true.

Thing is malfunctioning again and freezing.

Shit...

 

[Automaticman]

When it freezes are you getting BSOD's? Is there a stop code or driver mentioned? Are you sure the OC is stable? Have you tested it with Prime 95 - if it fails there it most likely rules out the GPU. If it runs Prime ok can you monitor GPU temp while in game?

 

[coolkingler1]

Well it was a bit of a false alarm, it was only 1 freeze and not in game.

Anyways, it works fine for now again.

No BSOD's nope. Nothing is mentioned, there is no overclocking on that CPU done.

But I will do some tests next time.

 

[Automaticman]

So when it was freezing in game before was it just crashing to desktop?

 

[coolkingler1]

So when it was freezing in game before was it just crashing to desktop?

Nope, just total freeze.