Unexplained and painful restarting problem in my AMD FX 8350 PC with Windows 8 Pro 64 bit

unitedopinions

Honorable
Apr 19, 2013
16
0
10,510
Hello everyone,

At this point I am totally confused, clueless, perplexed and feeling overwhelmed about what the issue is related to, so I've put it under 'systems'.

Since the last 5-6 months the computer has been restarting for no obvious reason. Early on the frequency was very low, so as to make it appear as a one off occurrence. Happened once a month or so. In the following months frequency started increasing and in the last 2 weeks it has been happening as a series. It restarts again anytime from the Windows login screen, to having opened a program. However, sometimes it stays on without restarting for 1-3 hours during which I can work.

This is the configuration of my system that I had got built in July 2013:

1)AMD FX 8350
2)ASUS M5A97 R2.0 motherboard
3)Sapphire AMD Radeon HD 6670 1 GB DDR3 Graphics Card
4)Corsair Vengeance DDR3 8 GB (2 x 4 GB) PC RAM
5)WD Caviar Blue 500 GB Desktop Internal Hard Drive (for secondary storage)
6)Kingston SSD SV300S37A/60G (For windows and programs)
7)Cooler Master Chassis Cabinet-Mid Tower-Elite 431 Plus-RC-431P-SATA X-Dock Case - with a total of 3 X 12 cm cooling fans
8)Corsair VS450 450 Watt PSU

I am running Windows 8 pro 64 bit with all the windows updates done regularly. I don't play any heavy games that'll take a toll on the graphic card. I play flash based online games, and sometimes very old games present on the pc.



What all I have tried so far:
1) When it first started, tried to look for a BSOD error code. In spite of error logging for crashes being enabled, no error is being logged. I tried installing 'WhoCrashed' and it too couldn't help as there were no logs present. As per suggestions I also ensured that a page file was present on the drive

2) When I googled for the pattern of restarts, it suggested updating graphic card drivers. I updated them to the most recent ones and since then AMD Catalyst service has been updating the drivers periodically. Also, after every restart I have been opening the chassis and touching all the components to see if anything, especially graphic card feels warm. Nothing feels even slightly warm.

3) I use Norton Antivirus and Malware Bytes and scan frequently. I scanned multiple times to check if it was a virus.

4) Updated BIOS and the chipset to the most recent version. It hasn't changed anything.

5) Some articles suggested it could be a faulty PSU. But I had got the wattage of the PSU verified from here at the time of building the PC, and I have checked it on the online power calculator again. Secondly, The system never restarts while in BIOS. It always does after going to the Windows screen.

6) One suggestion was to see if the heatsink was not clamped properly. I checked it. I didn't appear lose. Plus in BIOS the CPU temperature is always steady 44-46 C, what it always was.


Any thoughts? Please help.


 
Solution
My guess is the PSU. You say it doesn't restart while in the BIOS but that is the minimum load being put on the PSU - no hard drive access going on, minimal if any graphics, and minimal RAM access.

Another hint is that the frequency has increased. There's a lot more to a PSU than wattage. Components in the PSU heat up when under load and if it is for an extended period they can be damaged so that when they heat up again it takes less to cause the power shutdown. As heat goes up, resistance increases which raises the heat which causes the resistance to go up and soon you are into what's called a "runaway" situation.

The voltages can change under load and the PSU attempts to compensate for that. The PSU monitors its output voltages and...
My guess is the PSU. You say it doesn't restart while in the BIOS but that is the minimum load being put on the PSU - no hard drive access going on, minimal if any graphics, and minimal RAM access.

Another hint is that the frequency has increased. There's a lot more to a PSU than wattage. Components in the PSU heat up when under load and if it is for an extended period they can be damaged so that when they heat up again it takes less to cause the power shutdown. As heat goes up, resistance increases which raises the heat which causes the resistance to go up and soon you are into what's called a "runaway" situation.

The voltages can change under load and the PSU attempts to compensate for that. The PSU monitors its output voltages and when they go out of spec it shuts itself down rather than damage other components in the computer. That typically doesn't cause an error log because it is internal to the PSU which doesn't communicate with the OS.

You said your temperatures were OK but I'm guessing you are looking at readings reported through Windows which won't include anything for the PSU as far as I know. You could get a thermometer to try measure the heat on the components in the PSU but the cover would have to be off for that and that's fairly dangerous if you're not familiar with handling uncovered PSUs.

So the only way to test it is using voltage meters on the outputs from the PSU while it is under load and that gets rather complicated. You could take it to a shop but it would probably be cheaper to just replace the PSU.
 
Solution

unitedopinions

Honorable
Apr 19, 2013
16
0
10,510


Thank you for your responses @thx1138v2 and @tiny voices. I will have to try changing the PSU. But I am really very surprised to learn it's this. Because at the time of buying I had checked for brands in forums including here, and from the ones available here in India there was a discussion between Cooler Master and Corsair and most opined that Corsair is solid when it comes to PSU.

While I read this I have 3 questions come to my mind. Would be great if you could add your insights on this as well.

1) Are you saying that this particular piece I have is defective? Or are you saying that a Corsair 450 is not good enough to handle the configuration I have mentioned?
2) Do you think that during a power failure, a UPS with less than recommended KVA can stress the PSU and cause it to get damaged? I have a 900 KVA UPS that runs 2 computers with 450 watt PSU. I know that this is somewhat lesser than recommended, but can it potentially damage the PSU?
3) Any recommended brand / wattage for my configuration.

Many thanks for your answers again.