Question Computer seems to be dying; Can't find the cause

Myronazz

Distinguished
Sep 5, 2016
325
12
18,795
Hello... So yesterday I had the pleasure of troubleshooting my sister's computer for an ENTIRE day, and the entire morning of today too! (How fun!) Her computer seems to be dying but for the life of me I cannot figure what exactly is causing all these problems; is it the power supply? hard disk? RAM? I can't seem to figure it out so it would save me a WHOLE lot of time if I posted what I've been able to determine so far and get some opinions here as well, you have no idea how badly I need a second opinion right now. This is somewhat alot of information so i'll try to compress it down into bullet points so it's easy to read and digestible, i don't exactly feel like posting a wall of text and overwhelming people with info and bad structure

What is the problem?
Well... it seems that the computer freezes briefly, explorer will freeze for about 3 minutes and then unfreezes again. Sometimes the mouse freezes too and in even rarer scenarios the computer blue-screens with some weird errors I cannot find online. It's mostly a "Kernel inpage error" which apparently has something to do with either the hard disk or RAM. Some of you may go "Well okay there's your answer" but not really, i'll explain later on.

How is the freezing triggered?
This is the interesting part that makes things confusing. It started yesterday when she tried to install Fortnite, the Epic Games installer would attempt to install Fortnite and the computer would start freezing like CRAZY! It would even BSOD sometimes, you might think this is software related but something tells me that it's not.

What do I make of this?
Interestingly when it BSODs it would often run checkdisk on next boot and repair drive errors but although the drive may of have had some corruptions, checkdisk fixed them and SMART information is good, I also did a sector check just in case and it came out good, so the drive MUST be good, right? So that leaves RAM, I did a memtest86 and it gave RAM a pass with zero errors, so that's out of the window too. So what's next? Well I am not sure, is it the motherboard? Is the power supply dying? Well I doubt it's the power supply because it actually remains completely stable on a Prime95 Stresstest as shown below:

6BMmQv6.png


So... if the freezes were happening because the power supply was sending unstable power, now would be a good time to freeze but it didn't so therefore I don't think this is the problem

So what could it be?
I wasn't sure. But when I tried installing the latest NVIDIA drivers something interesting happened and so it gave me an idea of what the problem might be, the system started becoming unstable again and freezing all over the place when the installer was extracting the files... take a look in this screenshot:

cR1IwvS.png


This is a screenshot of when the system started becoming unstable again, which as I said was during the extraction of drivers, the system was freezing when the disk activity was full... and as a matter of fact, I was ONLY able to capture the screenshot when the system became slightly more stable again, which happened to be when the disk activity went down to 35% (see the task manager graph) so... coincidence? I think not... but then again I already said that the drive was fine, so how could this be?!?! Well perhaps it's not the drive itself, but maybe the SATA controller in the motherboard. This is the only logical conclusion I was able to deduct



This is all I have for now. For now I can only blame the motherboard. It could be a software issue but I don't really have any reason to believe that right now (maybe you can give me one?) I'll post a screenshot of HWMonitor for any people that want this information:
cEe3USt.png


If you guys can help me figure out what the exact problem is I would greatly appreciate it. Here are the full system specs:

CPU: Core 2 Quad Q9300
GPU: GTX 580
RAM: 2x4GB DDR3 @ 1066 MHz
Mobo: Biostar G41D3+
Power Supply: Evga W1 500W (Aware of the not so great quality but it's all we can do for now)
SSD: Weird Sasmung SSD that was probably pulled from a laptop @ 120GB

In case this turns out to be a bad motherboard, I can give her my H110M-R with a Core i3 6100 (Should be much better too) which I was planning to do eventually but didn't have any DDR4 sticks for it to actually work. Plus I want to be 100% sure it's the motherboard and not something else that might cause damage to the H110M-R once I install it, just being safe.
 
Last edited:

PC Tailor

Illustrious
Ambassador
Does the issue still occur in safe mode?
Do you have dump files of the BSOD you were encountering? or at the very least do you recall the error codes you were receiving?
Is the BIOS up to date?
Was anything changed before the problem started occuring?

I would remove all bar 1 RAM module for the duration of the testing. I'd also potentially be leaning to the storage drive being faulty - if CHKDSK was finding errors, it doesn't necessarily "fix" them - and if the errors are returning, then you'll be an endless loop.
 

Myronazz

Distinguished
Sep 5, 2016
325
12
18,795
I don't know if I have BSOD dump files. And even if I do aren't they quite large? Isn't also kind of unsafe to post the contents of your memory online? Or perhaps you are referring to something else? I already mentioned that the BSODs I was receiving mostly were something about "Kernel inpage error", there wasn't any hexadecimal code I could see, or maybe I missed it.

I will try doing what I want to do on safe mode, should of have thought of this before! And no nothing was changed before this

So i'll do what you suggested and post back, I'll also try running the computer with another, known-good Windows installation and see if this still happens
 

PC Tailor

Illustrious
Ambassador
No the dump files are usually configured as minidumps - which are the adverse, very small.
And it is not unsafe, we do it on here all the time. Myself and many others frequently assist in minidump debugging. Again you need specific software to be able to debug in the first place.

At the crooks of it, it would allow debuggers to see the system spec, what was running when it crashed, the nature of the crash, and where the crash took place in the system.

Quote often the kernel inpage error is storage drive related. Apologies - I accidentally didn't take that bit on as part of your original post!
 

Myronazz

Distinguished
Sep 5, 2016
325
12
18,795
Haha it's alright. I tried changing the SATA port of my SSD and at first, the HDD disk indicator was blinking, not the normal blink, it was indicating some kind of error, the interval between blinks was too perfect to be normal SDD use indication

I then changed it again to another port and this time no more blinking but as soon as the computer reached the Windows loading logo, the computer shut off... not sure why. I then tried again and it booted this time. Wtf was that about?

Currently I have booted from a Windows 8 Hard Disk and will try installing the same things that were making the computer freeze with SSD. And guess what, it went through extraction no problem, no freezing, nothing. It's installing just fine.

So the SSD must be faulty somehow? I But SMART info is showing good health and it has good sectors, so why...? Is this an issue with Windows maybe?

When the 'faulty' SSD booted again, Windows tried to repair it again! Seems like it's dying or something. But tests are showing it as healthy so why????

Do you think I should get a new SSD? I can get a 120GB one for £18 (Wow SSD prices sure are lower now) or can I do something to the current one to attempt to fix it? Something like a 0fill?
 
Last edited:

PC Tailor

Illustrious
Ambassador
So you tried a clean install and the problem has persisted?
Well the SMART attributes is only a summary, it doesn't catch everything. I've had plenty of drives appear fine on SMART attributes, but fail read/write tests, to find the drive was indeed faulty.

Do you have a picture of the SMART attributes you are reading? Personally I use HD Sentinel or CrystalDiskInfo.

If the freezing issue still occurs in safe mode, there is little doubt it is hardware - and I would lean to storage first based on the symptoms you've described. It can equally be the SATA cable or ports itself so it may be worth changing cables to see if the issue persists.
 

Myronazz

Distinguished
Sep 5, 2016
325
12
18,795
Well if all you say holds true, it's most likely a hardware issue, the drive still freezes on safe mode AND a totally different drive works fine and has zero freezing issues on the same cable and port. So this just has to be the drive that is dying, I can't seem to find another logical explanation. I did use CrystalDiskMark, it reported a 77% Good Health score, so yeah... we got the drive used so who knows what it went through before us. the SSD did pass a sector test though.

If the computer works absolutely fine with another drive I tried on it, I think it's enough evidence to conclude that the SSD is dying somehow, otherwise I don't know how else to verify that it's dying.
 

PC Tailor

Illustrious
Ambassador
77% wouldn't be good by my books.
This would tell me that whilst it theoretically is meeting the SMART thresholds, it is deteriorating.

And even one bad sector with irretrievable data can cause big issues. One of my last HDD failures was simply down to the PC sometimes freezing, but absolutely no windows updates successfully installing. I found a couple of bad sectors and tested another drive, and it worked perfectly.
 

Karadjgne

Titan
Ambassador
According to picture :
12v rail = 12.416v little high, but livable
5v+ rail = 5.081v spot on.
3.3v rail = 2.008v. Houston, we have a problem.

Well it's software, so that's a possibility when it comes to errors. So you'll need to do a manual check of outputs. Easiest way is with a digital multimeter set for DC.
Black = ground.
Red = 5v+
Orange = 3.3v
Yellow = 12v+
Check in the motherboard 20+4, don't even need to unplug, just shove the probe into the hole until it hits the pins. If the voltages do not align with what the wire is supposed to be, that's the issue. The Evga W1 is a basement budget psu, so unreliability is quite common.
 
  • Like
Reactions: PC Tailor

PC Tailor

Illustrious
Ambassador
That being said, @Karadjgne raises a good point that your 3.3 rail is going way out of a 5% tolerance anyway. Which can equally be causing a problem.
I still suspect that the issue you are trying to address is the storage drive, especially being as a different storage drive worked perfectly.

But the PSU could well be contributing as Karadjgne states. Regardless, if the voltages do not align - then you'll want to replace the PSU regardless as to what problem it is contributing to.
 
Last edited:

Karadjgne

Titan
Ambassador
Yeah, the motherboard uses 3.3v for most of the logic circuitry on the mobo, so has very little actual draw, but is still very important. 5v+ powers usb/Sata etc so that's the hdd's directly, but the logic circuitry once the data hits the mobo is 3.3v, so corruptions are entirely possible.

Which is why I suggest testing the pc with it actually running, just sitting at idle is fine. If the input voltages are good, the likelihood the issue is in the mobo goes up. Keep the meter in the 3.3v orange wire, then apply a load, like updating the drivers. Monitor the voltages. If it's still good, you have a bad mobo, somewhere a component is failing when subjected to heat/stress. If voltages plummet, well new psu is prolly a good start, but still no guarantee.
 
  • Like
Reactions: PC Tailor

Myronazz

Distinguished
Sep 5, 2016
325
12
18,795
Oh whoops... As I posted my previous reply for some reason @Karadjgne's comment didn't show hence why I didn't react

I actually did notice that the voltage was low on that rail, but didn't think much of it because no-one mentioned anything when I posted the thing so I went on to continue thinking nothing of it... Ugh so I do have faulty hardware? I do happen to have a multi meter so I guess I can check the rails on the power supply... so i'll do that and come back...
 

Myronazz

Distinguished
Sep 5, 2016
325
12
18,795
Well okay I'm done checking... I checked every orange wire in the ATX cable (and the SATA cable just to be safe) and it measures a good 3,4v (not 3,3 but I think that's within tolerance range) so that means the motherboard is bad...? But then why did the other drive work so well? I already ordered the SSD so we'll have to wait until it arrives, if it still throws issues then I guess the mobo really is bad...
 

Karadjgne

Titan
Ambassador
3.4v is ok. So at least you know now that that software is not always entirely accurate. Any reading could be suspect. Even though it's not a good psu, it seems to be working, so that's a plus, something eliminated. After you put in the new drive, you'll have something else eliminated. The more you eliminate as a potential cause, the narrower the focus.

There's 3 possibilities. Firmware, hardware, software. If you can eliminate hardware, that'd be great. Software is hard, but things like
  1. Press Windows Key + X to open Win + X menu and choose Command Prompt (Admin).
  2. When Command Prompt opens, enter sfc /scannow and press Enter.
  3. The repair process will now start. Don't close Command Prompt or interrupt the repair process.
can help, malware full scan, anti-virus full scan, use CCleaner (default settings), use CCleaner Registry tool (default settings, say yes to backup) etc can eliminate many potential issues.

Hopefully enough to allow the drivers and chipset drivers to be updated. I'd like to see a bios update too, but with freezing issues, I'd be wary.