1

I really can't figure out what is causing hard freezes on my system. They are very random, and not when playing games or doing some stress computational problems. But almost every time I run windows backup, after about 10-20 minutes I get hard freeze with a sound of whistle on speakers. (but also when not running backup, but those freezes are not as frequent, but still have same whistle sound).

It looks like is something on HDD. I have two HDDs installed, both Sata-2, AHCI mode. I've disabled the old one first, but it also came down to hard freezing up a system. Now I think that is possible that newer HDD is problem causer. Disk is Samsung F3 1TB. How can I check disk for possible errors?

Is there any possibility that the problem is somewhere else? I have updated all the drivers and system is pretty clean. I have no other problems whatsoever.

Thank's!

2
  • Have you run scandisk and checkdisk?
    – goblinbox
    Jun 27, 2010 at 1:58
  • Yeah, I've done Self-test utility in Ubuntu OS. No errors on both hard drives...
    – zigomir
    Jun 27, 2010 at 20:49

3 Answers 3

0

What can be said so far: it is either a driver or hardware problem. Cooling can also be an issue, you could monitor the core temperature to figure it out. Your motherboard software has probably a tool to show that figure. (but unlikely: you made some stress tests)

I also had quite a bad time with HDD drives on a brand new box: see sata: hdd stuttering on reads(led on). Because the box was new, I did not know where to start with. I tried almost everything out like switching from AHCI to IDE mode and many other games. Maybe this gives you some hints. Note that the tools to check the HDD didn't show up any problems in my case and with a replacement drive the problems were away. The disk just had a hard job in reading the data I guess, retrying for ages. If the symptoms are the same (HDD LED keeps being on during the freeze of more than a few minutes) maybe the same issue in your case.

4
  • Yeah, after a freeze HDD led is still blinking. And also, I've just checked whole drive with Ubuntu's disk check utility (extended self-test) and no errors were found. I really don't want to replace my drive yet... Stress tests were ok, temperature is stable...
    – zigomir
    Jun 27, 2010 at 20:40
  • I would try with another drive if you find or can get one from a friend for a short time. Was the same in my case, stresstests ok, HW tests ok, but drive was defect. After replacement with exactly the same model, it was fine.
    – jdehaan
    Jun 27, 2010 at 21:51
  • Blinking you said, in my case it was only always on, not blinking... The tests I made included some vendor tests (WD Lifeguard) and they also passed ok, SMART values were ok. Not sure what these tests are testing if they pass and the drive is broken :-(. I hope I do not push you into a wrong direction. Keep your mind open to other ideas...
    – jdehaan
    Jun 28, 2010 at 5:24
  • For now it looks like system is working. I've set CPU clock down to default and whole backup went through without freeze. Tomorrow I'll try again with over clocked CPU to see what is the point. It's interesting that when gaming or CPU stress testing there were no problems...Maybe just some power peaks didn't get through...
    – zigomir
    Jun 28, 2010 at 19:51
0

If it's not your hard disk, then it will be a driver or a hardware problem...

If it would be a driver, your best bet is to start a topic in the Sysinternals Forums - Troubleshooting to identify the driver by generating and analyzing a crash dump.

How can I check disk for possible errors?

Check out my recent post:
When to stop using a hard drive? What rules/software apply?

2
  • A crash dump would be nice but you don't get any on freezes, only if you encounter the BSOD...
    – jdehaan
    Jun 27, 2010 at 10:36
  • That's not true, even if your computer is frozen you can still generate a BSOD through CrashOnCtrlScroll and check with WinDBG where the problem lies. Another option would to be to remotely connect to the Kernel over a Firewire or USB Debug Cable and then check the call stack of the active process. But for more details he should ask on those forums for support or read about CrashOnCtrlScroll and BSOD analysis himself... Jun 27, 2010 at 12:11
-1

So, my problem was over-clocked CPU and more than that, CPU voltage was set lower than it's default value. But the strange thing is, that with that setting CPU stress tests and playing games worked fine.

You must log in to answer this question.

Not the answer you're looking for? Browse other questions tagged .