Since mid-January, my wife’s PC is becoming increasingly erratic. Computer boots up OK, and then has about 5-10 minutes before it will freeze up for half a second, and then go through the first couple seconds of a reboot. Screen goes black and the video card info flashes on the screen (as it does in a normal boot). Instead of hearing hard drive sounds (set as the 1st boot device), the CD-ROM drive (set as 2nd boot device) spins up, and then the system seems to lock up. I have to restart the whole thing, when I get another 10 minutes to run some test to figure out what’s wrong.
Specs: Athlon II X3 435 CPU, ASRock motherboard, 4GB DDR3-1600 memory, GeForce GTX 460 videocard, Hitachi 1 TB hard drive, Thermaltake 500W power supply, Windows 7 Ultimate 64-bit operating system. CPU is 2 years old (I think), all other components 1 year old except the power supply, which is brand new.
What I’ve tried so far:
Replaced power supply; old one was 400W, came with the case (Codegen), and was a few years old. Voltages reported in the BIOS for the new power supply all look good.
Checked all power connections are securely connected to internal components. Over and over and over.
Replaced video card. Didn’t think this was causing the reboots, but since I was upgrading the power supply anyway, I ordered a new card for my machine & gave my wife my old one (the GTX 460). Old card in wife’s machine was a Radeon HD 4830.
Ran chkdsk on hard drive; didn’t seem to have major problems – no bad sectors, etc. It reported some boot sector info was incorrectly labeled as unallocated, but that this had been repaired.
Ran Windows memory diagnostic; no problems.
My next thought was to reformat the hard drive and do a complete reinstall of Windows, but I’m really just flailing about. I have no idea what’s wrong and would welcome any theories or suggestions.
Install a temperature monitoring program like speedfan and see if the operating temperatures are climbing during those 5-10 minutes.
What cooler do you have on there?
You could also try booting up a Linux live CD and let it run for a few hours. If that doesn’t lock up, then it’s more likely to be a Windows issue (I think it’s unlikely, but…)
Also check for motherboard BIOS updates, if the patch notes for any of them indicate they might fix an issue you have then go for that…
Yup, use Memtest86, preferrably from a boot disc and allow it a couple runs through the tests. Your description of symptoms sounds a lot like a memory issue. Consider turning off the default setting of “Automatically Restart Windows After Error” as this may help you diagnose what’s happening. That setting is under “Startup and Recovery” in the Advanced System Settings in Control Panel. Essentially you’ll see a blue screen the next time it crashes after unchecking this setting, and the blue screen should help you determine if you’ve got a memory error occuring or something more serious.
If that comes out OK, the next culprit might be the CPU heatsink and fan. If you are comfortable doing so, unhook the heatsink assembly and remove the fan. Clean it out and clean out the grooves on the heatsink as well. Dirty/dusty environments tend to fill the fan and heatsink with layers of caked on dust, and that lowers it’s efficiency a great deal, sometimes to the point where it no longer draws enough heat away from the CPU. You could also attempt to extract and reseat the CPU, and if you’re feeling really industrious you could even redo the seal between the CPU and the heatsink (I suggest Arctic Silver for that job).
Thanks for the tips, folks. No resolution of the problem yet, but updates from last time:
I created a Memtest86 CD & the machine passed three runs with no memory errors.
I reinstalled the older video card, a Radeon HD 5770, rather than a Radeon HD 4830 as I originally reported. Although it didn’t fix the reboot problem, I was having trouble with driver updates on the newer NVIDIA card anyway.
I blew out all the dust & crud from around the CPU heatsink/fan (there wasn’t that much). Using Speedfan temperature monitoring, there don’t seem to be any heatspikes preceding the lockups. CPU temps stay between 36 and 37 C, and core temp varies between 33-34 C.
From your comments above, it sounds as though a hard drive reformat/Windows reinstall is unlikely to fix the problem, & I’d rather not go through the hassle. Tomorrow I’m going to reassemble the system from scratch, on the off-chance there’s some screw or something causing a short somewhere. Assuming the problem still continues, I figure I need to replace either the motherboard or the CPU (or both).
I’m almost positive both the CPU & motherboard are each two years old. Any recommendations on which one to replace first? I figure I’ve only got a 50-50 chance of being right by guessing, so if I can tip the odds to something like 65-35, I’ll feel better. I’ve had one motherboard failure & no CPU failures in my past systems, but the motherboard failure had completely different symptoms than this – the whole machine just wouldn’t boot.
OK, turned off the Auto Restart Windows as folks suggested. When the system crashed, it simply locked up at the Windows Desktop. I had just finished checking to make sure the auto restart button was unchecked.
So there is no BSOD with an error code. The system is just frozen with a few windows open on Desktop.
Need to go to dinner now; I’m going to try to check the MEMORY.DMP file when I get home to see if it says anything useful. Thanks for the continued help, folks.
It really sounds like, since it’s happening a few minutes in, it’s mobo related / heat related. Something is heating up enough to sever/split a contact/lead.
If you want to test the CPU itself use Prime95. I tried to google it quickly but their main site is dead, and too lazy to figure out where to find the latest version. Someone might know of a replacement to that but that’s what I’ve always used when OCing when I wanted to test out CPU stability.
Checking this tonight. Tried to read the Memory.Dmp file, but kept saying access is denied, which is weird, because there’s only one account on the whole machine. So root account should have Administrative access.
Update: Machine ran in safe mode without hanging all night. Does this suggest there’s some driver corruption rather than a CPU or mobo problem? And I still can’t access the Memory.Dmp file.
Have the machine performe some tasks in safe mode, the more intensive the better, to see if it hangs. Had machine that also run fine in safe mode and even after Windows reinstall hangs continued turned out that safe mode didn’t stress the machine enough to cause the hangs, faulty motherboard.