PC restarts suddenly

TobyD · Sep 3, 2014

Executive Summary:
-Experiencing spontaneous restarts (immediate shut down), with occasional blue screen.
-Often (but not always) happens on graphics/memory intensive task, e.g.:
*Clicking links in Chrome
*Adjusting display resolution/configuration
-Usually get Error 41 in Event Log (usually with 0X0 for all parameters).
-I suspect the power supply, but don't know how to confirm/troubleshoot due to the intermittent nature of the problem.

System/Hardware:
Windows 7 Pro, 64 bit
Motherboard: Gigabyte GA-P55-USB3
CPU: Intel Core i5 760
Graphics card: Radeon HD 5850
Power supply: Antec Basiq Power 550W Continuous Power
Ram: 2X 4GB Skill RipJaw. (F3-10666CL9D-8GBRL)
Hard Disk: 1 TB

Description:
For several weeks, I've been battling a problem where my computer suddenly restarts (shuts down immediately, sometimes after showing a blue screen warning for a split second.) This first started happening after installing Windows updates, and for a while I could fix the problem by restoring my system to a point prior to the updates. (Uninstalling them didn't seem to work). After this stopped fixing the problem, I noticed that my graphics card driver was for the 5700 series instead of the 5800 series. After I installed the correct driver, the problem disappeared for about a week. When it started happening again, I disabled the AMD driver alltogether. That bought me another few days, but then the computer did a sudden restart even with that driver disabled (and, in that most recent instance, without any click from me).

Triggers include:
Opening new Chrome tab (happened often)
Pressing "next" in gmail (in Chrome). (happened often)
Selecting "Make this my main display" and clicking "apply". (once/twice?)
Something in Solidworks. (once)
Login (a few times)
Passively reading a notepad document (latest instance).
Copying ~380 MB from flash drive to hard drive. (Restarted just after initiation.) (Transfer completed no problem after restart) [EDIT]
Ejecting flash drive (after creating bootable USB for memtest86) [EDIT]

Based on the research I've done on this site and others, it seems like my power supply could be
struggling. If this seems like it might be the problem, is there any way to confirm/troubleshoot before investing in a new one? (And if I do need to just buy a new one, any recommendations?)
Thanks!

EDIT:
Check CPU temps at crash. They were around 40 deg. C.
Windows Memory Checker (OK)
memtest86 in progress.

Also, here's a log I've been keeping of a few of the instances:
8.20.2014 Note: Sometimes get brief (1/4 sec) stripey screen before blue screen.

8.18.2014 Note: Working after restore prior to Windows updates. Action center says to "Install your video card driver" (Recommending 5800 series. Autodetect installed 5700 series.)

2014.08.22 Using 5800 series driver for graphics seems to have fixed problem. (I'm stupid.)

8.27.2014 Sudden restart when logging in in the morning (after restart?)
Then sudden restart when clicking on email in Gmail (Chrome)
Note: Failed Windows update attempt this morning and yesterday morning: KB2942664
Error Code 800B0100 (Missing or corrupt file).
Restoring to yesterday; running update auto-fixer thing.
Restarted during/after restore. Upon login, got messages: "System successfully restored to 8/27/2014, 9:32 AM" (later than above restarts, but I selected that restore point because the next one back was 8/22), and "System recovered from an unexpected shutdown."
Sudden restart when moving Notepad window between screens.
Restoring to 8/22/2014. "Scan for affected programs" says only Windows security update will be affected, but I know I've installed at least some MATLAB toolboxes since then (8/26).
Restore successful. (Have to reactivate Windows)
Still sudden restarts after restoring to 8/22. (Computer worked well from 8/22 to 8/27.) Graphics driver is 5800 series (correct.)
Disabled Graphics driver (Radeon HD 5800 series). No restarts.
(Activated Windows again.)

9.2.2014 Sudden restart while reading a notepad document. (Didn't click on anything.) First time since disabling AMD graphics card driver.

Checklist:
Safe Mode: Works
Normal Mode w/ only Monitor 1 (Dell with speaker) (in upper slot): Failed. (Straight reboot upon clicking Google+ photo.)
Normal Mode w/ only Monitor 2 (Dell, no speaker/USB) (in lower slot): Failed. (Straight reboot upon clicking craigslist link.)
All new (8/20/2014) Windows updates uninstalled. Restarted: (Note: KB2952664 reinstalled itself?)

Windows Memory checker reports no problems.
Newegg power supply calculator estimates 484 Watts. (Have 550 W power supply)

Usually get Error 41 that looks like:
- EventData

BugcheckCode 0
BugcheckParameter1 0x0
BugcheckParameter2 0x0
BugcheckParameter3 0x0
BugcheckParameter4 0x0
SleepInProgress false
PowerButtonTimestamp 0

But sometimes it looks like this:
- EventData

BugcheckCode 292
BugcheckParameter1 0x0
BugcheckParameter2 0xfffffa8008089028
BugcheckParameter3 0xf2000000
BugcheckParameter4 0x20008f
SleepInProgress false
PowerButtonTimestamp 0

Shain Taylor · Sep 3, 2014

Hi, it sounds like either RAM or your psu. Are your CPU temps okay? Which core I5 is it? If it's a haswell/devils canyon then your PSU needs to support the new C-states

jitpublisher · Sep 3, 2014

Download and run memtest, with each stick of memory by its self to eliminate or pin point a bad memory stick. Random crashes, blue screens, random errors you are describing are consistent with a bad memory stick.

TobyD · Sep 3, 2014

Shain Taylor :

Thanks for the input Shain.

The CPU is an i5 760, which I don't think has any special low-power states.*

I downloaded a CPU temp monitor. I'll set that to log and then try to trigger the problem by re-enabling my big graphics card. (It would seem odd to me if temps are the problem, because the restart is usually so immediate after a click. But I'll let you know.)

*A bit of background: This computer was built by a colleague back in 2011/12, worked fine, and then sat idle for a couple years until I bought a new hard drive and took it over. So the hardware is a little older than most builds.

TobyD · Sep 3, 2014

jitpublisher :

Thanks. Windows memory checker reported no problems, but I get the impression that that's not nearly as thorough(?). I'll run memtest and report back.

TobyD · Sep 3, 2014

Shain Taylor :

CPU temps are at ~40 deg. C before the last crash, so that doesn't seem to be the problem :\

Shain Taylor · Sep 3, 2014

TobyD :

Strange. Could be a driver issue. You could try reseating your CPU? Let us know what memtest says...

TobyD · Sep 4, 2014

Shain Taylor :

I ran memtest86 overnight with both sticks of ram in place. Errors started popping up after the second pass. I've cleaned the sticks and slots and I'm now running the test with one stick at a time. I'll post the results.

Shain Taylor · Sep 4, 2014

Okay atleast we've sourced the problem!

Shain Taylor · Sep 4, 2014

TobyD :

It's strange, I have the exact same RAM as you and have had no issues... Are you sure the timings are correct in the BIOS? They should be 9-9-9-24 at 1.5v...

TobyD · Sep 5, 2014

Shain Taylor :

OK, big update:
As I reported previously, I ran memtest86 (not 86+) with both RAM sticks in: Got errors after 3rd(?) pass. (Seemed to be increasing with each pass.)
I then cleaned the RAM stick contacts with alcohol and the slots with compressed air.
Ran memtest86 with each stick individually. (Each in their original slots (1 and 3). Got 0 errors after 4 and 18 passes, respectively.
I checked my timing: It's 9-9-9-24 at 1.5V.
I reset all BIOS options to "Load-optimized defaults"
Re-enabled Radeon graphics card driver, started Windows and tried to trigger a crash...and I haven't been able to!
So, I'm going to rerun memtest 86 with both RAM sticks together over the weekend. Maybe it was as stupid as dust in the RAM slots??

I'm cautiously optimistic about this, and I really appreciate the help so far! I'll post an update after the weekend.

Shain Taylor · Sep 5, 2014

TobyD :

Shain Taylor :

OK, big update:
As I reported previously, I ran memtest86 (not 86+) with both RAM sticks in: Got errors after 3rd(?) pass. (Seemed to be increasing with each pass.)
I then cleaned the RAM stick contacts with alcohol and the slots with compressed air.
Ran memtest86 with each stick individually. (Each in their original slots (1 and 3). Got 0 errors after 4 and 18 passes, respectively.
I checked my timing: It's 9-9-9-24 at 1.5V.
I reset all BIOS options to "Load-optimized defaults"
Re-enabled Radeon graphics card driver, started Windows and tried to trigger a crash...and I haven't been able to!
So, I'm going to rerun memtest 86 with both RAM sticks together over the weekend. Maybe it was as stupid as dust in the RAM slots??

I'm cautiously optimistic about this, and I really appreciate the help so far! I'll post an update after the weekend.

Awesome

TobyD · Sep 8, 2014

Shain Taylor :

So...bad news and good:

Bad news: After running all weekend (38 passes), with both sticks of RAM installed, memtest86 logged 44,851 errors. (I can provide details if that would help.) This is after each stick tested clean individually. I've found varying opinions online about whether this situation means that the RAM is bad or not.

The good news: None of the usual triggers are crashing my computer. So...I guess I'll just go from here unless it starts crashing again. (Unless anyone has more thoughts/advice.)

Big thanks to Shain Taylor for the help!

Shain Taylor · Sep 8, 2014

TobyD :

Shain Taylor :

So...bad news and good:

Bad news: After running all weekend (38 passes), with both sticks of RAM installed, memtest86 logged 44,851 errors. (I can provide details if that would help.) This is after each stick tested clean individually. I've found varying opinions online about whether this situation means that the RAM is bad or not.

The good news: None of the usual triggers are crashing my computer. So...I guess I'll just go from here unless it starts crashing again. (Unless anyone has more thoughts/advice.)

Big thanks to Shain Taylor for the help!

No problem! Could you select a solution? You could go into the BIOS as see if the timings are correct when both sticks are installed

TobyD · Sep 8, 2014

Shain Taylor :

TobyD :

No problem! Could you select a solution? You could go into the BIOS as see if the timings are correct when both sticks are installed

Yep--as far as I can tell, cleaning the RAM sticks/slots was the answer, so I've selected your first solution which suggested faulty RAM.

(And yes, the timing is still 9-9-9-24 with both sticks installed, so that isn't/wasn't the problem.)

Thanks again!

TobyD · Sep 9, 2014

A day later, it's begun rebooting itself again. Sigh...

To recap:
-memtest86 shows no errors when testing RAM sticks individually.
-I do get errors when running memtest with both RAM sticks installed. (I'm still confused about this.)
-Timing and voltage seem to be correct in BIOS. (9-9-9-24 at 1.5V).
-I've reset the BIOS to "Load-optimized defaults". (Should I try flashing the BIOS?)

Shain Taylor · Sep 9, 2014

TobyD :

Strange. Were they bought as a dual channel set? are you running them in dual channel?

Alec Mowat · Sep 10, 2014

Shain Taylor :

They should be dual channel as long as they are in the correct slots.

TobyD · Sep 10, 2014

Alec Mowat :

Yep, they're in slots 1 and 3, as per the MOBO manual's instructions. (I can't find anything in the BIOS that says Dual Channel, but it does list all 8GB of memory.)

Alec Mowat · Sep 10, 2014

TobyD :

It's irrelevant

Replace your PSU yet? As soon as I saw "For several weeks, I've been battling a problem where my computer suddenly restarts ", I realized it was a PSU problem. Buy a good, high quality Bronze certified 550W PSU and you'll be fine. Any random shutdowns are generally PSU related. Don't bother troubleshooting until you replace your PSU at this point. That PSU installed is pretty cheap anyway.

TobyD · Sep 11, 2014

Alec Mowat :

Thanks Alec. Yeah, the PSU seems like a reasonable culprit...does it explain why memtest fails when both sticks of RAM were installed (but showed no errors when testing individually)?

As for PSUs, does this one seem like a decent choice (or am I being too cheap)?:
http://

johnbl · Sep 11, 2014

bugcheck 292 = 0x124 WHEA_UNCORRECTABLE_ERROR

common overclocking/ overeating bugcheck/ wrong voltage or clock

you can use the windows debugger and dump the reason that the CPU called the bugcheck
use !errrec on the address of the WHEA structure.

I would turn off hardware acceleration in the browser on the assumption that the video card is pulling too much power from the PCI bus and causing voltage fluctuations to your CPU.

Alec Mowat · Sep 11, 2014

TobyD :

That one is fine for your setup.

It's probably getting power fluctuations and since RAM is volatile, it drops the signal and fails. If your system is randomly shutting down though (not BSOD), that's an instant power related warning. You can check your voltages on some motherboards to be safe, but I'd just replace it and go from there. It's cheaper than buying new RAM.

You may want to buy a better PSU than what you currently have. If it's not the problem, at least you come out with a better PSU in the end.
http://www.amazon.com/Corsair-Series-Modular-Bronze-ATX12V/dp/B00ALYOPSS/ref=pd_tcs_subst_pc_2?ie=UTF8&refRID=07SK1GVEPSVW6P1VPNFD

Alec Mowat · Sep 11, 2014

johnbl :

I don't recommend this. It's just a quick fix, the power issue still exists and you will still hit these errors with any other application using the videocard. This is something my dad would do.

johnbl · Sep 12, 2014

- Dumping the memory structure in windbg is not a fix, it helps to determine the cause of the problem.
- Turning off the browser hardware acceleration is also not a fix but helps to isolate the true cause of the problem
and provides a short term workaround.

Alec Mowat :

PC restarts suddenly

Reputable

Honorable

Splendid

Reputable

Reputable

Reputable

Honorable

Reputable

Honorable

Honorable

Reputable

Honorable

Reputable

Honorable

Reputable

Reputable

Honorable

Honorable

Reputable

Honorable

Reputable

Polypheme

Honorable

Honorable

Polypheme

Share this page