PC either hangs or reboots constantly

[artman70]

I have a home built PC with the following configuration:

CPU - Intel Core 2 Duo
Motherboard - ECS P965T-A
RAM - 4 1 gb DDR2 667
SATA hard drive (500gb)
NVidia GeForce 7600 GS
Running Windows 7
PC is about three years old.

When booting windows, or when running MemTestx86 from the USB drive, my PC will either lock up or restart randomly. Usually, when the system is doing something (copying a file, loading a web page, etc). This occurs both in safe mode and outside of safe mode.

I have tried the following to resolve the problem:

Replaced thermal paste and reseated the CPU and fan
Moved the graphics card from one PCI-E slot to another.
Replaced the PSU
Tried swapping out RAM to see if one might be bad.
Ran memtestx86 from a USB drive to test the memory.

If I leave the machine in the BIOS, it stays up. If I boot into MemTest or windows, it will eventually crash. If I try to update the graphics drivers it crashes will copying files.

Have even tried taking one memory stick and trying it in different slots to see if I had a bad slot on the motherboard.

Any thoughts on how I can confirm whether the motherboard or the CPU is the culprit (or something else, for that matter)?

 

[Raidur]

You used a completely different set of for sure working RAM?

Can you see your CPU/NB temps in BIOS?

 

[__Miguel_]

From my experience, that seems like a memory-related issue.

Just to be sure, though, do a "bare-bones" boot: disconnect everything non-essential (while checking everything is working in the process), and boot with only mobo+CPU+PSU+HDD+GPU and one stick of RAM (if you have more than one, try every single one of them in each slot, to filter out any defective RAM slot - or the chipset failing on you, I had an Asus P965 board I had to exchange for a new one because the chipset simply died...).

Since your system is not new, there might be dust accumulation somewhere. While you're preparing for the bare-bones tryouts, inspect every surface and slot, just to make sure dust (or other foreign element) isn't causing some sort of bad contact.

If you have already tried all that, please test the RAM modules one by one on another motherboard. Also, Raidur's option is good, try another memory kit.

If I understood correctly, memtest is also freezing the PC, not only Windows, which usually rules out a corrupt Windows install.

You might try a new CPU, if available, but as I said before, that sort of problems usually means a bad memory stick, a bad memory slot or a chipset defect. Excessive temperatures don't usually cause issues when only taxing the memory subsystem...

Hope this helps.

Miguel

 

[artman70]

Well, I only have the four sticks that I had installed. I did install each one seperately and ran memtest. It crashed each time. I even tried using different slots to see if it was a bad slot. I would have a hard time believing that all four of my memory cards went south at the same time, but I suppose its possible.

CPU temps are around 28 C in BIOS and 35 C in Windows.

A little history:

The problem first occurred about a week ago. My PC suddenly rebooted itself, and each time rebooted more and more swiftly. After trying to clean out dust, things seemed to get worse. At one point, the PC would not even boot into windows, rebooting before the Starting Windows message, or sometimes even during the initial memory test. It would even restart in BIOS sometimes. Core temperature was normal, and all the fans were turning.

At this point, I tried each memory stick individually, to no avail. I also tried reseating the graphics card. After doing some research, I bought a new PSU and installed it, but nothing changed.

Then, I tried moving the graphics card to another PCI-E slot, and the computer was able to boot. It was at this point that it started either locking or rebooting, usually but not always when copying files (I was moving some important files to a USB drive). I even tried safe mode, but with the same behavior.

Then, I tried resetting CMOS, and it was in BIOS that I noticed that the core temperature was running around 60 C. My best guess is that all my tinkering must have knocked the heat sink assembly loose. I removed that and the CPU, removed the thermal paste and reapplied, reinstalling the CPU and heatsink assembly. No change.

I then tried Memtest on several combinations of memory, but each time, the machine would reboot or hang before the test could complete. No errors were found before the reboot.

Here are my current PC health statistics (while in BIOS) if that helps

CPU Temperature - 24C
System Temperature - 25C
CPU Fan Speed - 1670 RPM
CPU Vcore - 1.28V
NB Vcore - 1.23V
5VSB - 5.05V
+12V - 12.03V
+3.3V - 3.36V
VBAT - 3.31V

Thanks for any guidance. I have run out of ideas, other than replacing everything.

 

[Raidur]

I would try some known working Memory. I haven't come across a failed Memtest with good RAM, however if I did, I would assume Motherboard failure.

Wait for a second opinion before purchasing a new mobo however.

Also, clarifying. When you reseated your heatsink, your CPU temp went back down to normal, correct?

 

[artman70]

Yes. The temps I posted are the current temps

 

[Raidur]

Another clarification, Memtest freezes your computer? If so, does it show errors before doing so?

*edit ah nvm I found the answer to that.*

 

[__Miguel_]

Yep, either test those memory sticks on a known good motherboard or a known good set of memory sticks on that motherboard.

However, anything short of a massive power surge should NOT be able to fry four memory sticks at once... I've had one (of two) die on me following a short-circuit, but even then it only started corrupting data... Intel CPUs, on the other hand, are known to be able to work perfectly fine even with burnt or incorrectly-bridged pads, so I wouldn't put my money on the CPU.

That seems more of a chipset failure, but as Raidur said, test out the memory sticks (and slots) first.

Is it just me to suddently feel like the P965 chipset was prone to spontaneous death?

Miguel

 

[BuddyW]

Check for Bubbled up or blown capacitors on the Motherboard and Video Card. If either item has a capacitor problem that is most likely the problem. If they both look OK, I would replace the power supply.

 

[__Miguel_]

Buddy, he has already changed the PSU. Two duds in a row is not very likely, since the problem stays the same, right?

Miguel

 

[Raidur]

Is your NorthBridge very hot to the touch? If so think about replacing/remounting that heatsink as well.

I'm really thinking its your Motherboard. Especially if you're still getting problems with everything disconnected besides the CPU/RAM/GPU.

It's either that or the CPU assuming you don't have a short somewhere.

I'd still think about trying a different kit of RAM. I know it is HIGHLY unlikely that is the issue due to you having 4 sticks to try, but it is a small possibility.

Your slowly growing BSODs could have been due to your RAM slowly failing. You could have still been boot-able until now because your first stick or 2 weren't as bad as the last 2 down the chain, until now.

This is unlikely, but if there is any possibility of you getting a hold of another stick of RAM, or another PC for you to test your sticks in. I'd do so before going out and buying a new motherboard.

If not I'd go for the mobo idea. I believe Best Buy will let you return things (not sure about motherboards, but they will memory I believe). Usually I'm not one for this kind of diagnosing, however Best Buy is one of those titan companies that makes millions ripping people off on certain products. (some of their things are decently priced however)

I'm guessing it's a bad mobo. :/ Good luck mate. Still stick around for other opinions.

Also, I read somewhere once that USB Legacy can cause Memtest to fail so some people disable that while running it. However that won't solve your entire problem just being able to pass a Memtest, obviously.

 

[artman70]

Northbridge is not very hot.

I think I will try some new RAM.

As far as the BSOD's, I never once got the blue screen. Its always been a hard freeze, or a sudden reboot.

I use Frys, they have a 30 day return policy, and have never had an issue with returning stuff (in fact, I returned the PSU I bought when it didn't fix the problem)

As far as the USB Legacy option...I couldn't find that option on my BIOS. Not sure if its supported on my mobo.

Thanks again for everyones help. I'll let you know tomorrow the results of the new RAM.

 

[artman70]

I simply cannot believe that all my RAM went bad at once.

I bought 1 2gb DIMM and plugged it in this evening. So far so good. I've been running fine ever since. Installed new video drivers (the machine would crash if I tried it before), played some video, surfed the web, let it sit a little, hibernate, resume, etc. So far everything seems fine.

This will remain a mystery, because I have never heard of all the RAM going bad at once (unless, I suppose, lightning hit, but that should fry just about everything, and the problem started on a sunny day.

Thanks again for all your suggestions. I really appreciate it.

 

[__Miguel_]

Hmm, weird indeed.

However, RAM is the single most sensible piece of hardware in any computer. Even the HDDs are like tanks, compared to the RAM, so it is conceivable an energy spike (caused by lightning or otherwise) fried the whole lot. Odd, but possible.

Probably a good idea to invest in a UPS sometime down the line, no?

Anyway, glad to know you're OK(ish) now. Good luck.

Miguel

 

[artman70]

Quick update: I bought a second 2 gb DIMM (to get the system back up to 4 which is what it had before all of this began) and the problem re-occured. So, I moved the new DIMM to another slot, and the problem went away.

So, it sounds like I had one or more bad slots on the motherboard ( as well as some bad RAM, because I tested from every slot with my old RAM when I was originally troubleshooting the problem.

I suppose its always possible that all my jiggling around inside the box may have exacerbated the problem.

In any case, thanks again to everyone who chimed in. I hope this can help others, should they ever stumble upon a similar problem.

 

[artman70]

Best answer selected by artman70.

 

[__Miguel_]

Damn, bad RAM AND at least one bad memory slot. That is some serious bad luck there...

Glad to know it's fixed now, though.

Oh, and thanks for selecting my answer as the best one. People sure are friendly in these forums.

Take care.

Miguel