PC Randomly Freezes. [Solved]

[icewolf461]

Hi. My PC randomly freezes and the only way to power it off is via the PSU switch. I did SFC scans and swapped out the old RAM with new one but it still freezes. I also scanned for viruses and ran CCleaner. I don't get any errors at all, it just freezes. I'm completely stumped, any help would be appreciated. Thanks.

PC Specs:
GPU: EVGA GTX 650 1GB
CPU: Intel E5400 @ 2.7Ghz
RAM: 2.50 GB
Motherboard: ASROCK G31M-GS R2.0
PSU: ThermalTake TR2 600W
HDD : WD 160 GB
OS: Windows 8.1 Pro

Nothing is overclocked.

 

[chusteczka]

Can you provide two screenshots of CPUID HWMonitor? One screenshot at idle and the next under load for 5 minutes. Prime95 can put the system under a load. I am wondering about temperatures and PSU voltages.

I will mention now that I do not like ThermalTake products and wonder if the PSU is the culprit. ThermalTake products tend to be manufactured cheaply with poor engineering. They used to have good cpu heatsinks though.

Do you have an AVR or battery backup for power line conditioning?
Have any windows updates failed recently?
Can the system pass two passes of Memtest86+?

Motherboard: ASROCK G31M-GS (DDR2, Skt LGA 775)

 

[Ketchup]

ah, Wolfdale. Pretty good CPU. Still have one running at my parents' house.

Looking forward to some results based on a nice post by chusteczka, the only thing I would ask is what is it freezing on: sitting on the desktop? heavy load? working on the computer? leaving it alone?

Temps and voltages should be a good help here.

 

[icewolf461]

Sorry I didn't post the pics yet my PC keeps freezing on me. It freezes idle, gaming, on desktop, boot splash screen, while using media player. Everywhere...but random. I don't think I have an AVR or battery backup for power line conditioning. No Windows updates have failed recently or at all and I haven't tried memtest yet. I will try to before it freezes again.

 

[chusteczka]

You have an older system, so it appears safe to assume that it was running fine before now. Therefore, your electrical lines providing power are probably fine. If it is a power issue, then it would be the PSU. Yet it cannot hurt to get an Automatic Voltage Regulator (AVR). Once I got one, many erratic issues disappeared. I now often give them as gifts.

Newegg.com - OPTI-UPS SS1200 6 Outlets Voltage Stabilizer
Newegg.com - OPTI-UPS SS1200 Voltage Regulator (600 watts)
Newegg.com - OPTI-UPS SS2000 6 Outlets Home AVR Series Automatic Voltage Regulator (1000 watts)


The freezing happening often enough to impede installing HWMonitor and getting screenshots makes the problem pretty bad. Results from HWMonitor may prove useful.

Did anything change recently that might have created the freezing?
Do you have another spare power supply that can be used to test your system? Pull out the old, put in the new or different to test?

 

[Ketchup]

Are there other computers in the house/office and are they running fine?
If you plug your computer into other jacks in the home/office, do the crashes still occur?

If you answered yes to both, the problem is not outside your computer. I would like to see temps/voltages before making recommendations of what you should buy.

CPUID makes software I like for this:

http://www.majorgeeks.com/files/details/hwmonitor.html

 

[Burpo]

Hi. My PC randomly freezes and the only way to power it off is via the PSU switch. I did SFC scans and swapped out the old RAM with new one but it still freezes. I also scanned for viruses and ran CCleaner. I don't get any errors at all, it just freezes. I'm completely stumped, any help would be appreciated. Thanks.

PC Specs:
GPU: EVGA GTX 650 1GB
CPU: Intel E5400 @ 2.7Ghz
RAM: 2.50 GB
Motherboard: ASROCK G31M-GS R2.0
PSU: ThermalTake TR2 600W
HDD : WD 160 GB
OS: Windows 8.1 Pro

Nothing is overclocked.

Nobody noticed the 2.5Gb of RAM?
Mismatched memory

 

[Ketchup]

Nobody noticed the 2.5Gb of RAM?
Mismatched memory

We didn't miss it, it's just that that running different sticks alone would not cause a PC to freeze, and judging by age has probably running fine with that configuration for a while.

 

[chusteczka]

Nobody noticed the 2.5Gb of RAM?
Mismatched memory

Saw it but figured the only problem would be slower than dual channel would otherwise provide. I do not think mismatched memory would cause stability issues.

A recommendation for faster memory would be to obtain (2x4GB) DDR2-800 memory through eBay. Prices on that memory are super high. I would honestly prefer to build a new DDR3-1600 (or 1866) system rather than purchase more DDR2-800 memory.

And Ketchup is correct, I should not be so cavalier in recommending an AVR purchase. The majority of the homes around me were built in the 60's and 70's where line conditioning really helps clean up the power going into the computer. The OP's issues are clearly not caused by unclean power supplied from the walls and my recommendation for an AVR is superfluous.

 

[icewolf461]

Idle: http://oi62.tinypic.com/wuqjia.jpg

Heavy Load (Prime95): http://oi62.tinypic.com/34imtd2.jpg

No new changes recently. I do have an old PSU but the cables are too short. This is the only PC in the house. I never tried another outlet, they're all plugged into a surge protector.

 

[chusteczka]

The voltages from the PSU are fine. Your ThermalTake power supply appears to be working fine.

CPU and Video temps are fine. The load temps on CPUTIN and AUXTIN are fine. However the Max temps for CPUTIN and AUXTIN are too high and have me worried, at 92C and 95C. I expect those are for the northbridge or southbridge chipset temps.

The Northbridge chipset controls data through the memory and video card.
The Southbridge chipset controls data to and from the hard drive(s) and USB ports.

It is possible the Northbridge and Southbridge chipsets are running hot due to dried thermal paste under the heatsink. It may help to pull your board out, carefully remove the heatsinks for the north and southbridge chipsets, and replace the thermal paste. If you do, be careful not to break the plastic pins that hold the metal heatsink in place. You will be needing them again. I like to use needlenose pliers to squeeze them together and push them through from the backside.

The chipset heatsinks are shown in the below image in the red bordered rectangles.

1. Does your system seem to freeze when the hard drive is being heavily accessed, such as during defragmentation or a virus scan? This would affect the southbridge. It would help to run a manual virus scan while leaving the HWMonitor open to watch the temp.
2. Can you do the Prime95 test again but with the setting that heavily tests the memory? This will test the northbridge. It would help to leave the HWMonitor open to watch the temp.

Other than chipset temperatures due to dried thermal paste, I am not sure what might be happening to your system. Whatever is happening is happening intermittently but frequently. Such as a virus scan accessing the hard drive and overheating the southbridge chipset when the system is idle.

 

[icewolf461]

1. Yes, freezes when it requires HDD activity. When it freezes, the indicator is either always lit, or off; never blinking.

 

[chusteczka]

Now we are getting somewhere. The motherboard is a few years old, replacing the heatsink thermal paste could be seen as a necessity for routine maintenance. Or you could verify by watching the temps while running the two tests for Prime95 memory and then a manual virus scan.

I expect you will soon be pulling the board to replace the thermal paste for the two chipset heatsinks.

EDIT: I also expect that your system would freeze if you attempted to test your memory with Memtest86.

 

[icewolf461]

1. Manual virus scan: http://oi61.tinypic.com/2wcegx5.jpg

2. Heavy memory test (Prime95): http://oi62.tinypic.com/kf1p5e.jpg

 

[chusteczka]

Temps are normal. I hope you gave it about 6-10 minutes of test time. I expected your system would have frozen with either of those two tests. At this point I am stumped. It might still help to pull the chipset heatsinks and replace the thermal paste but I am not sure what the cause is right now.

 

[icewolf461]

I was on another forum yesterday, and a guy said to raise the DRAM and vCore voltage by +0.05 and see. The vCore option was locked but I was able to raise the DRAM voltage. (Not by +0.05 manually, it had auto and other pre-set voltages. http://imgur.com/xXDrVEv) So he said try 1.8 and I just did, so far no freezing.

 

[chusteczka]

support.microsoft.com - Windows 8.1, Windows 8 or Windows 7 hangs or freezes
Microsoft mentions it useful to update the system drivers.

wikihow.com - How to Stop a Windows Computer from Freezing
WikiHow mentions to clean up the computer. If all else fails, then format and reinstall Windows.

tomshardware.com - computer randomly freezes
This guy mentions to reinstall the OS.

tomshardware.com - Computer Freezes randomly
This thread mentions drivers possibly crashing.


At this point, I do not think it is physical. Except it couldn't hurt to replace the chipset thermal paste.
Google seems to point to software.
https://www.google.com/?gws_rd=ssl#q=computer+freezes+randomly

 

[icewolf461]

I tried to reinstall Windows but it froze on setup screen.

 

[chusteczka]

I was on another forum yesterday, and a guy said to raise the DRAM and vCore voltage by +0.05 and see. The vCore option was locked but I was able to raise the DRAM voltage. (Not by +0.05 manually, it had auto and other pre-set voltages. http://imgur.com/xXDrVEv) So he said try 1.8 and I just did, so far no freezing.

I did not consider memory voltages since you mentioned no overclocked bios modifications. But that could definitely be it. With DDR2 memory, you are safe between 1.8v - 2.0v, even up to 2.2v as the bios settings suggest. Sometimes changing from AUTO to the standard setting (1.8v) helps force the motherboard to hold that voltage setting.

Good job, I am glad that is helping. If it freezes again, it would be safe to bump the memory voltage up to 1.85v and even 1.9v if necessary. Let's hope it holds.


EDIT:

I tried to reinstall Windows but it froze on setup screen.

Then it would not be software and points towards the memory voltage. The memory voltage could be the VIN5 that is slipping down to 1.7v on HWMonitor.

 

[icewolf461]

Hope:

 

[chusteczka]

I am sorry that I did not think of the memory voltage but glad you found the help elsewhere.

 

[icewolf461]

Damn, it still froze. Took a little bit longer though.

 

[chusteczka]

Bump the memory voltage up another 0.05v, to 1.85v or 1.9v. I really wouldn't go above 1.9v or 1.95v.

Run Memtest, I realized that is the test result that would have earlier brought my thoughts to memory voltage. Although at this point Prime95, testing heavy on the memory, would do just as well.

Other possibilities include:

1. bumping cpu voltage,
2. replacing the PSU,
3. and getting an AVR to stabilize the supplied power. An Automatic Voltage Regulator protects against voltage spikes as well as momentarily strengthens the voltage during a momentary brownout.

EDIT: From your bios image, try increasing the other voltages, very slightly in this order.

1. +1.5v voltage
2. NB voltage
3. VTT voltage

 

[icewolf461]

Okay, I increased them up a notch.

 

[Ketchup]

I see you all have been busy (last night in my world).

icewolf461, I think you are doing a great job here. The temps and voltages look good to me.

But from what I can see, your problem looks to be your RAM or power supply. You may also want to use it as a sign that it's time to upgrade if you can.

I wish you could have checked for minidumps before wiping the drive. Would have been nice it at least one made it in.

On the RAM, don't forget to test the one stick at a time during the install (not sure how many slicks you have).

If we assume there were no memory dump, this would be a sign of power issues or overheating. Since you are not overheating, might give a little nudge to replacing the power supply first. Possibly try a line conditioner if there is any doubt about that.