[SOLVED!!!] Bizarre random freezes - Tried everything I could think of

Toggle sidebar Toggle sidebar
E

eduardowoj

Commendable
Jun 12, 2016
29
0
1,540
FOR THOSE WHO ARE GETTING THE SAME SYMPTOMS, HERE IS WHAT SOLVED THE PROBLEM FOR ME: It was the wifi driver. Seriously, just that. I have a TP-LINK wifi card, and this company does not provide a Win8+ driver, only for Win7 and under. So, when I installed the card, Windows automagically installed Microsoft's driver, which is faulty. It was a matter of uninstalling Microsoft's driver and installing TP-LINK's Win7 driver in compatibility mode. The rest of the thread details what a faulty wifi card can do on a computer, it's scary.

Hello guys, this will be quite a long post. In short, I'll say what's my problem on the first paragraph, what are my specs on the second, and I'll follow detailing every single test I did to try to locate what's giving me problems.

TL;DR version: Gaming PC, custom built, randomly freezes. Truly random, sometimes when browsing internet, sometimes when gaming, sometimes when watching something, but never when it's idle. When it freezes, the audio also freeze (the speakers go "prrrrrrrrrrrrrr" for a second) and then mutes.

Specs:
- MB: Asus Sabertooth 990FX R2.0
- CPU: AMD FX8320E
- RAM: Kingston HyperX Fury 16 GB (2x8GB, 1866MHz)
- PSU: Corsair RM650x
- GPU: Asus GTX970 OC Turbo II
- Cooling: Corsair H60 hydro cooler

I had posted my problem before, and marked as "solved" when the problems lifted for a while (giving me this illusion), so instead of reopening the old post I'm writing this one.

So, let's go to what I already tried.

1) Fiddled with CPU multiplier on BIOS to fix the RAM frequency on 1866 MHz (the mobo autodetects my sticks as 1151 MHz)
2) Increased RAM voltage from 1.5 V to 1.55V and then to 1.6 V, as suggested for those HyperX, to try to make it more stable
3) Returned to auto configurations
4) Tried fitting the RAM sticks on different slots
5) Tried fitting and refitting the sticks several times
6) Tried different RAM chips, a Corsair Vengeance 16GB (2x8 GB, 1866 MHz) from another computer that is stable
7) Tried running with just one RAM stick (and with just the other too)
8) Disabled Turbo Cores on BIOS
9) Resetted the BIOS
10) Cleared the CMOS and resetted the BIOS just in case
11) Reflashed the BIOS
12) Set memory control to 2T instead of 1T
13) Disabled antivirus (Avira AntiVir PRO) while playing, ¯\_(ツ)_/¯
14) Uninstalled Windows 10 and went back to 8.1
15) Reinstalled everything because it kept freezing

To test the memories (tests 1 to 7), I ran memtest86+ overnight and it found no errors. I did this on both RAM sets. So I ruled out memory problems. To check for CPU problems, I ran PRIME95 on the most cruel torture setting available and the computer crunched numbers for more than 7 hours without freezing a single time (worth noting that CPU temps kept cool, thanks to the water cooler).

With this, I was tempted to rule out CPU problems also, but on the next day, less than 30 minutes of Skyrim were enough to make the computer freeze. So I began suspecting of the VGA. However, 5 hours of FurMark later, with no hiccups, and I had to accept my VGA was working as intended.

Just in case, to check the hardware as a whole, I booted up StressLinux and left it running for almost a day, something like 16 or 17 hours. Not. A. Single. Freezing.

At this point, I pretty much didn't believed in hardware problems anymore. Then I remembered that, when the PC was upgrading to Windows 10, I had energy problems and it had shut down while upgrading. I tried running SFC /SCANNOW and DISM and it found problems, but I couldn't fix them. So I just reinstalled Windows (and kept in 8.1, because I had had some problems with Win10 before) and the problems seemed to vanish. This was the point where I closed the last thread.

Fast forward to mid-July, and the problems began to appear again. Albeit less often than before, but the PC keeps freezing every now and then. Mainly during gaming, but sometimes while browsing the interwebz or watching something.

I need help. I don't know what to test anymore, and what to check. I'm trying, starting yesterday, to disable the antivirus every time I play, but it could take some time to feel safe that the computer wouldn't hang anymore. Any more ideas of what (and how) can I test? Any theories of what's happening with this computer?

Thanks in advance for the awesome support!
 
Sort by date Sort by votes
cub_fanatic

cub_fanatic

Honorable
Nov 21, 2012
1,005
1
11,960
Did you ever test your HDD/SSD? I like to use HD Sentinel (http://www.hdsentinel.com/) The free version will give you a pretty good idea about the health of your main drive. It sounds like you've ruled out the usual suspects like RAM, PSU and GPU.

Also, how are you connecting your PC to your monitor? HDMI, DVI, VGA or DP? I once had a bad HDMI cable that would cause problems like frozen picture, loss of sound and artifacts.
 
Upvote 0 Downvote
A

avarice

Distinguished
May 10, 2006
633
0
19,160
I would normally suggest a simplify/test/isolate approach - but your issues seem to occur during specific software and none of the burn in programs you used. The only untested commonality that I can think of between the gaming, browsing and watching 'something' - is internet access. However that should not cause issues - just slowness.

You COULD try to reset everything back to default (BIOS/ect.) perform a comprehensive virus/malware scan and then see if this comes back.

I know this might not be much help, so good luck.
 
Upvote 0 Downvote
E

eduardowoj

Commendable
Jun 12, 2016
29
0
1,540


In the test I ran in StressLinux, I included heavy HDD testing. This test included writing, copying, moving and deleting big files, or several small files at a time. Since the computer didn't hanged during this long test, I ruled out HDD too.

The computer was hooked to the display via a VGA cable, with a DVI adaptor. But I got a new DVI cable to connect it. Since the VGA cable was what I was using on other PC, I know it was good. I tried connecting the display to a different DVI on the video card too. :(
 
Upvote 0 Downvote
Faux_Grey

Faux_Grey

Honorable
Sep 1, 2012
747
1
11,360
What does windows event manager say every time it hangs?

I'd recommend you turn of automatic restart on system failure so you can actually see what the error message is when your computer crashes.

Nice howto:

http://www.howtogeek.com/howto/windows-vista/help-troubleshoot-the-blue-screen-of-death-by-preventing-automatic-reboot/


If it happens again, jot down the BSOD code.
 
Upvote 0 Downvote
E

eduardowoj

Commendable
Jun 12, 2016
29
0
1,540


Yes, since I pretty much ruled out hardware issues, software is what's left. I already did a deep virus scan some time after reinstalled everything, and my computer came out clean. I even checked for rootkits and checked the MBR.

I'm a bit suspicious of the antivirus, but it would take time to check, since those hang ups are stochastic.

 
Upvote 0 Downvote
E

eduardowoj

Commendable
Jun 12, 2016
29
0
1,540


I tried verifying Windows logs, but those are hard freezings that I'm getting. Windows don't dump information into the logs because the whole computer freezes, instead of being a Windows instability that just hangs the computer.
 
Upvote 0 Downvote
MeanMachine41

MeanMachine41

Judicious
May 8, 2014
5,195
336
33,390
Hi eduardowoj :)

Quite an extensive report and thanks for that. It seems you have tried just about every trouble shoot regarding your DIMMs and GPU but that may not be the issue. Prime95 is a harsh tester and some swear by it for stability checking, however I don't cause its too harsh and does not test other sub systems. Also good you have ruled out your DIMMS being an issue using Memtest86+.

Lets do some stress testing my way and see if you have an instability in your system Bios settings.
Download AIDA64 and HWMonitor to your desktop and put them side by side.
Run the tests in AIDA64 starting with CPU and FPU. You can check your sub systems later.
Run the test for 10mins and stop the test if temps reach 80C.
Take screen shots at the 10min mark and link the results here. You can use IMIGUR as file host and link the BB url here.

HWMonitor should be showing all rail voltages, that is +12V, +5V, +3.3V. and 100% utilization on all cores. Reason for this is I want to see if your PSU is doing its job under load. I also want to check for voltage droop under load. These can be rectified by Tweaking Bios.

Do not OC the system and if you have then reset CMOS back to default settings. Also what Bios version are you running.? If not the latest then please update. If your not sure how to do this correctly and safely then I will detail the steps for your MB.
OCing should only be done in Bios in case you have used third party software for any reason.

Once we have a baseline picture of how the system is performing under load at stock frequency then we can begin.
 
Upvote 0 Downvote
E

eduardowoj

Commendable
Jun 12, 2016
29
0
1,540


Thanks, I didn't thought about AIDA64. I'll run these tests this night, or tomorrow's morning. My BIOS isn't OC'd, and it's currently running with all default options. Also, I'm running the latest release from ASUS.

As soon as I get the results, I'll post them here.
 
Upvote 0 Downvote
E

eduardowoj

Commendable
Jun 12, 2016
29
0
1,540


Hi MeanMachine41!

I did the tests, and it it seems that everything's fine. The PC didn't froze, the voltages stayed uniform, the watercooler did an amazing job. The only thing that didn't worked was IMGUR, so I uploaded the screenshot in another platform:

stresstest.png


(direct link if that doesn't work)

I forgot to add to my original post that my PSU is a Corsair RM650x, of which I've read excellent reviews.

What can I do now?
 
Upvote 0 Downvote
MeanMachine41

MeanMachine41

Judicious
May 8, 2014
5,195
336
33,390
Ok the test on CPU, FPU, and Rail voltages under load are fine. What is not right is the Core temperature readings which can't be correct as it should be at least 55C-65C. It could be HWMonitor is reading incorrectly so just try using AIDA64 and extend the tests for 1hr with all boxes checked and if OK then its within Windows.
You stated earlier that sfc /scannow found problems and could not fix them. If your system boots up into Windows without issue you may still have a corrupt system file. What were the energy problems you had when upgrading windows.?
I had a file corruption issue in Windows 10 and the only way I could fix it was to do a clean install. If your virus system has ruled out the possibility of malware, and Device Manager indicates that all drivers are up to date and no yellow triangles then that and all the testing you have done in between, then I'm out of ideas.

So far you have run tests using Synthetic algorithms. There is a program "Realbench" which uses realworld apps to stress your system and one thing I have found is that if you can run the tests and pass the benchmark then your system is stable. Try it.
 
Upvote 0 Downvote
E

eduardowoj

Commendable
Jun 12, 2016
29
0
1,540


I think 55C-65C would be too much, because HWMON idle temperatures are consistent with BIOS measurements. Also, I have other computer with a FX8350, but the same Corsair Hydro Series H60, plenty of room inside the chassis and fans positioned to maximize air flow that have the same range of temperatures, even in heavy loads.

Some time ago I recall reading the difference between the two measurements on HWMON is because AMD puts a thermal circuit inside the die, while the Asus MOBO has also a CPU meter close to the slot. So one is the measurement of the die-center temperature, while the other is the temperature of the slot. Finally, my other computer's CPU temps barely reach 35C after hours of gaming, with quite similar specs, so I wouldn't expect such high temperatures for my 8320E...

As for the sfc /scannow, that was before I reinstalled everything. What happened was that, on the first time, I installed Windows 8, which I have the license, and began the procedures to upgrade it to Win10. After it downloaded some 150 updates, when it was installing them, we had an energy spike and the computer shut down. We waited one hour and a half, turned on the PC again and began installing the updates one more time, but another energy spike made it shut down yet again. On the next day we finally managed to upgrade all and update it to Win10.

But since I was having problems and neither sfc nor DISM could run, I decided to do a clean install. This is the installation I'm running now.

I'll run the full 1hr AIDA64 test, and post the results when it's done. But I'm suspecting of something here... Video card. Ok, I did a FurMark test before, but it only seemed to me like a benchmark. What I noticed is that this computer can be idle for days, without a single hiccup. Every time it froze, the screen wasn't static. What I mean is that the video card was processing something: Either a game, a video, or scrolling the page.

Is there a program other than FurMark to stress test the GPU? AIDA64 can do that? I want to double check its health...

UPDATE: AIDA64's full test includes GPU stressing. I ran it for 1:05 hrs, and here are the results. (and direct link, with full size). By now, I think what's left is try to play games with Avira AntiVir turned off... Maybe it is somehow interfering with the video driver...

stresstest2.png


UPDATE 2: With Avira turned off, the computer froze again after almost two hours of Skyrim. This isn't a problem with Skyrim per se, because it also froze in other occasions with Wolfenstein, Fallout, Rocket League, Pinbalk FX 2, Touhou and etc...

I'm officially out of ideas again.
 
Upvote 0 Downvote
MeanMachine41

MeanMachine41

Judicious
May 8, 2014
5,195
336
33,390
All the games you have listed have one thing in common. They are all available on Steam. Are you using Steam.? I wonder if somehow your Steam client is the cause and would a re-install help.?
I have read others having issue with steam updates. I don't know, Just grasping at straws.
Might be worth submitting a ticket and see what they say.
Also: Have you tried the file verification process.?
Go here: https://support.steampowered.com/kb_article.php?ref=2037-QEUH-3335
 
Upvote 0 Downvote
E

eduardowoj

Commendable
Jun 12, 2016
29
0
1,540


Thanks for the effort, but it's not a problem only with steam... It froze a couple of times with Netflix, some others with youtube, and other times playing games outside steam. Besides, this happens now and on the previous installation.

I guess I'll try to swap this VGA with the graphics card from another computer, if I somehow manage to squeeze this monster Asus card into the cramped case of the other.

Do you know if there is a bootable VGA stress test tool? I would like to double check the card's sanity.
 
Upvote 0 Downvote
E

eduardowoj

Commendable
Jun 12, 2016
29
0
1,540


No problems!! I'm grateful for the time you spent trying to help me ;)
 
Upvote 0 Downvote
MeanMachine41

MeanMachine41

Judicious
May 8, 2014
5,195
336
33,390


Your welcome eduardowoj and good luck. If you do happen to find the culprit, please report your findings.
 
Upvote 0 Downvote
E

eduardowoj

Commendable
Jun 12, 2016
29
0
1,540


I just noticed some things this morning... I tried to play Wolfenstein: New Order, which is a pretty intensive game on the hardware. The PC froze twice, the second time less than a minute after I begin playing.

So I got some GPU stress test softwares. I downloaded GPUTest, which includes several different tests, and ran 7 tests simultaneously, each on testing a different thing. I also monitored the temps, and the GPU reached 80C with 100% load. But, 20 minutes later, it was still working as a breeze.

Funny how it hanged way quicker with a heavy game than it does with Skyrim and Fallout 3, older games.

So I opened my mind for other ideas and, apparently, many people experience lockups with the same wifi adapter, a TP Link 300 mbps N adapter (sorry, don't have the exactly code because I'm on mobile now while fiancee plays Skyrim wih no hangups until now). I'll try to follow the steps those people followed to solve their problems.

Also, just in case I'll get some hardware monitor software to log ALL THE DATA it can, so I can plot graphs of each information to try to grasp some information.

Finally, I think I'll give OCCT a shot. I'm somewhat suspicious of my Corsair RM650x...
 
Upvote 0 Downvote
A

avarice

Distinguished
May 10, 2006
633
0
19,160
You could simply go out and purchase a long network cable, or move your PC close enough to your router to test with an existing cable.

I have not gamed on a wireless for ages - gigabit Ethernet is the way to go.
 
Upvote 0 Downvote
E

eduardowoj

Commendable
Jun 12, 2016
29
0
1,540


Unfortunately, as much as I want, I can't hook this PC to the network with a cable, because it's quite far from the switch/router...
 
Upvote 0 Downvote
E

eduardowoj

Commendable
Jun 12, 2016
29
0
1,540
Funny thing, I'm almost marking this thread as solved. I'm now highly suspicious of the TP-LINK wifi card. Let me give some examples:

Playing Skyrim with internet on: Freezes in less than half an hour
Playing Skyrim with internet off: 6 hours of continuous gaming
Playing Wolfenstein with internet on: Freezes in less than 5 minutes (it froze in 30 seconds with me once!)
Playing Wolfenstein with internet off: 2 hours non stop that could be more if I didn't had so much motion sickness with FPS

So I followed some suggestions I found and installed TP-LINK's driver for Win 7 in compatibility mode, replacing the Microsoft-provided driver. Not only the connection is more stable, but it haven't hanged up a single time since then.

I'll keep testing, but if it doesn't freezes anymore, I'll mark as solved and write the solution on the first postage.
 
Upvote 0 Downvote
E

eduardowoj

Commendable
Jun 12, 2016
29
0
1,540
I'm closing the thread because I found out what was happening. It was precisely the wifi card driver.

TP-LINK doesn't provide an official driver for Windows 8 or 10, only for 7 and under. So, Windows automatically installed Microsoft's driver, which is faulty. When I downloaded TP-LINK's Win7 driver and installed it on compatibility mode, the problems were gone. I'm almost a month without a single hiccup on this computer.

Thanks for all the effort and for all the help!
 
Upvote 0 Downvote
MeanMachine41

MeanMachine41

Judicious
May 8, 2014
5,195
336
33,390


Thanks for reporting back and sharing your results eduardowoj :)
I'm sure many searching with similar issues will find your thread helpful.
Glad you solved the problem.
 
Upvote 0 Downvote
You must log in or register to reply here.

TRENDING THREADS

Latest posts

Moderators online

Share this page

COMPANY
RESOURCES
FOLLOW
Top Bottom