Desktop computer freezes, then is unable to reboot

Everything was working fine until last week. I turned it on to copy some files, Windows gave a garbled blue screen and after hitting the reset button it wouldn't even get to the BIOS POST. Turned it off, waited some time, then tried again and it booted, but some minutes later the same thing happened.

I'm guessing one or more of the components could be failing, but I wouldn't know where to start.

• processor - they're supposed to last 10 years in average, this one will turn 13 next month; temperature never goes above 50°C so it's not overheating
• RAM - three sticks the same age as the processor; haven't been able to run a memtest to find defects
• power supply - its wattage far exceeds the computer's maximum load, but it's 7 years old and had to be repaired 2 years ago
• video card - replaced it with a new one 7 years ago, haven't done any 3D gaming or heavy graphics work for years, I even underclocked it by around 50% to reduce power consumption and make it last longer
• hard disks - managed to run a quick SMART tests on both and they came out OK, but who knows?

My first thought would be security shutdown due to temperature. Especially if the computer refuse to boot imediatly after the shutdown, but can be rebooted after some time to cool down.

• Are you sure that the sensor on the CPU is ok? What if you get your finger on it, does it melt?
• Same thing for the GPU temperature? What if you put your hand close to the gpu ship.

When it shutdowns try to identify if a component (event the power supply) seems to be hot.

If you can't find any sign of overheating, try to leave only one RAM stick at a time and check all 3 of these. Maybe one of your stick is close to die, it happend to me with 3 almost new sticks.

Quote:

---

processor - they're supposed to last 10 years in average, this one will turn 13 next month; temperature never goes above 50°C so it's not overheating

---

I have a CPU built in 1996, so I wouldn't even suspect it.

Quote:

---

RAM - three sticks the same age as the processor; haven't been able to run a memtest to find defects

---

You need to do a through test with either MemTest86 or Memtest86+. Short tests of less than 1 hour do not find anything. And I would test them one by one (plug in one stick, and test, then take out the stick and place the other one and test).

Quote:

---

power supply - its wattage far exceeds the computer's maximum load, but it's 7 years old and had to be repaired 2 years ago

---

PSUs do not cause blue screens. As for power requirements, take the following example: Your computer takes 300W in normal operation, on BIOS POST it'll take 350W - 400W because it checks all components in a relatively short time, making it operate in full load. If it did not give enough power, it would not even be able to POST.

Quote:

---

video card - replaced it with a new one 7 years ago, haven't done any 3D gaming or heavy graphics work for years, I even underclocked it by around 50% to reduce power consumption and make it last longer

---

It could be the GPU. You need to test this too.

Quote:

---

hard disks - managed to run a quick SMART tests on both and they came out OK, but who knows?

---

Sometimes SMART logs are not a good way to test if or not a HDD is OK. Try to power it on a different computer, and place your hand on it, while it is operating. See if you can feel weird bumps, powerful vibrations, scratching sounds, or any other anomalies.


I suspect is either RAM or GPU or ... it could be the motherboard. If all else fail, get yourself a bottle of isopropyl alcohol and take out the motherboard and start rubbing with q-tips/or paint brush the entire surface (top and bottom) to make sure all static electricity is removed. The alcohol itself is non-conductive.

I would aim for the memory sticks first, in my experience these are the most problematic. You said you have 3 sticks, try to take them one by one and do some tests, maybe you can isolate the faulty one(s) if that's the case.

Secondly I would check for the CPU cooling, is the fan too dirty? Try to monitor cpu temperatures from bios first since your windows is shaky.

Quote:

---

Originally Posted by anthony-joal

Are you sure that the sensor on the CPU is ok? What if you get your finger on it, does it melt?

---

I hope it doesn't, I can't regenerate fingers :wtinfoil:

Thanks for all the advice, I'll get on this as soon as I have enough time. I made a thorough "maintenance check" of the computer case a few months ago, complete with cleaning up, replacing both the CPU cooler and thermal paste for better ones and rewiring everything to get the best airflow, so I'm ruling dirt and overheating out... but you never know.

Honestly, since I use my laptop almost exclusively and will be moving out soon, I'm tempted to just take out the hard disks, keyboard and mouse and leave the rest behind. But I've had this thing since forever, so I'll do my best to get it working if possible.

if not RAM, probably bad motherboard capacitors.

True, swollen capacitors can cause these "random" failures. They were the reason the power supply had to be repaired.

So, I spent some hours yesterday trying to figure this out.

First, I replaced the power supply with a known working one. After doing this, the computer wouldn't boot at all (not even for a few minutes as it did before). A CMOS wipe and a thorough cleanup of everything, including the 20-pin power connector as well as RAM, CPU and AGP slots, made it start up once more.

I then checked the temperature as Anthony suggested, and it seemed in line with what the sensors said, definitely not hot enough to shut down. I also cleaned up the processor and reapplied the thermal paste, even though I'd done it a few months back, just in case :unsure:

I checked the motherboard capacitors. None of them looked swollen or damaged.

The thing was booting up and seemed to work fine now. Encouraged by this, I put everything back in, closed the computer case, and used it for around two hours. Nothing bad happened, until I got a blue screen related to nv4_disp.dll when I was ready to turn it off for the day :wbaeh:

I gave it a second try today. Blue screens over and over, different reasons each time (page fault in nonpaged area, attempting to write to protected memory, critical system process terminated). I was ready to blame the video card, but decided to restore an old system backup just to be sure. Booting into Mini Windows XP resulted in yet another blue screen caused by ntfs.sys this time, and running memtest86+ afterwards revealed tons of RAM errors. In hindsight I should have checked the sticks one by one as was suggested (which I still need to do to find which the faulty one/s), but I guess I finally found the cause.

Quote:

---

Originally Posted by anon

blue screen related to nv4_disp.dll ,,.... but I guess I finally found the cause.

---

If you haven't found a fix yet you can check these sources out:-
Steps To Resolve The Nv4_disp.dll Error Information & Resources To Fix Common Computer Problems
How To Fix Nv4_disp.dll is Missing / Not Found Error
https://answers.microsoft.com/en-us/...6d24a63?auth=1

I'm not sure but the "blue screen related to nv4_disp.dll" is a supposedly driver problem (according to various sources on Google). Perhaps you can change your drivers to see whether that is causing the problem (do keep a backup incase this problem spurs on something else). However, if you can't start your pc, you may need to reinstall windows.

As for the drivers, I mainly rely on DriverPack Solution to solve my problems. However, you should confirm the installation of the correct drivers by the checking out the original site.

This video card has been discontinued, the driver I'm using is the last official one. I was ready to blame the card, until the blue screens continued both in safe mode and Mini XP (neither of which load the "real" driver). Mind you, that doesn't necessarily mean the problem lies elsewhere, I'm just much more willing to blame faulty RAM after memtest86+ found thousands of errors.

Unfortunately I have no replacements for either piece of hardware at hand to try and be sure.

M just here to laugh on your misery.. A big ass " Ha Ha Ha"

Quote:

---

Originally Posted by hellman

M just here to laugh on your misery.. A big ass " Ha Ha Ha"

---

200 Don't care

After testing all the RAM sticks individually (and even on every slot, just to be completely sure), it turns out the one in the second slot was bad. I removed it and everything works fine. The replacement costs just $2.50, I'll buy it this week to be done with this ordeal =]

Finally got a replacement stick, memtested it individually, no errors.

During the process, I noticed the timings on it as well as one of the existing ones were 3-4-4-8, while the other was 2.5-4-4-8. I set the memory timings in the BIOS to Manual instead of By SPD as they were before and forced the CAS latency to 2.5. This boosted the transfer speed to 972 to 1000 MB/s, with no apparent degradation in stability after a second round of testing. I wanted to lower the CL all the way down to 2 and overclock memory by a few dozen MHz to see what the boost would be, but ultimately decided against it.

So, this is now fixed, and to top it off performance should be slightly better than it was before =]