Random BSODs and applications crashes on a fresh Win7 install

[R4ndom]

Yesterday I took everything off and reinstalled - the cards, fans, RAMs, CPU (removed thermal component, applied new, resat, etc) and other connectors - gpu, MB, etc.

I left BIOS on default, except for memory timings, because for some reason BIOS sets it to 1T instead of 2T, and timings 9-9-9-24-33 instead of 9-9-9-20-33. I didn't change voltages, everything went fine. I ran prime95 for 6 hours straight and it detected no errors.


Today I start the PC, prime95 again and it spits out Rounding was 0.5, expected less than 0.4 ... prime froze my PC, so I had to reboot manually. So, I go to BIOS and I see that MB again set the command rate to 1T! I changed it back to 2T, reset the other timings, bumped CPU voltage to 1.45 instead of 1.40 standard, vddnb to 1.350 and I'm running prime again.

I also had a couple of BSODS these days, but won't upload them, as in the meantime I reset/resat everything. Anyway, this time it was "memory management" errors as reported by WhoCrashed, not whea anymore.

You mentioned that CPU idle power may be too low - what would be the remedy to that - I already changed the pwoer plan in windows to balanced, as when set to high it results in a very annoying high-pitched tone coming from around CPU-PSU area.

As for apps crashing, well, basically everything crashes and what's strange, this is very ofthen NOT logged by the Windows event manager (which has also crashed these days). It just crashes, but is not followed by the usual windows message: the programme has stopped and it gives you the choice of restarting it, closing, looking for help online.

At some point it was firefox which would fail every now and then, but among apps listed by you, apart from Fx, there's call of duty modern warfare - IW3MP, sopcast - but this hasn't crashed for a really long time, Prime, irfanview - also haven't seen any crash in weeks, windows explorer, recently I get quite often the following error: .NET Runtime Optimization Service (clr_optimization_v4.0.30319_64) - Service reached limit of transient errors. Will shut down. Last error returned from Service Manager: 0x8000ffff.

There was also an appcrash called livekernel event, which seems to be connected with GPU - but this was also before I resat the GPU.

Now among other apps crashing there's also skype, open hardware monitor, svchost.exe_SysMain, error side by side with a code 72, and there are also crashes of SCII, without any particular reason, but especially when I quite the application. Then MsMpEng.exe, svchost.exe_wuauserv, WinFlash.exe, PnPDriverImportError.

In Windows event viewer under windows logs/system I can see a multitude of errors (sometimes several times every second) connected with Service Control Manager with the following text: The service Computer Browser depends on the service Workstation, which cannot be started due to the followin error: the dependency service doesn't exist or it has been marked for deletion. Then this line is followed by another one: The service Workstation depends on the following service: MRxSmb20. This service cannot be installed.


As for the error reporting text, it's all the same in the phrases you quoted, so I'll translate the first, as the only thing that changes is the app names.

2012-04-21 11:33 Application Hang Program firefox.exe w wersji 11.0.0.4454 zatrzymał interakcję z systemem Windows i został zamknięty. Aby zobaczyć, czy jest dostępnych więcej informacji dotyczących tego problemu, sprawdź historię problemu w panelu sterowania Centrum akcji.
Identyfikator procesu: c88
Godzina rozpoczęcia: 01cd1fa6ff9f4216
Godzina zakończenia: 18
Ścieżka aplikacji: D:\Mozilla Firefox\firefox.exe
Identyfikator raportu: ce425251-8ba5-11e1-a87b-14dae9c5e52a

Programme firefox.exe version 11.0.0.4454 (I now run 12.0) stopped interacting with windows and was closed. To see if there's more information available concerning this problem, please check the problem history in the Problem Reports and Solutions control panel. 
Process ID: c88
Start time: 01cd1fa6ff9f4216
Termination time: 18
Application path: D:\Mozilla Firefox\firefox.exe
Report ID: ce425251-8ba5-11e1-a87b-14dae9c5e52a

Thanks again for your time, for the chronicle BIOS is now set as follows and running Prime95:

CPU v = 1.45
RAM v = 1.65
vddnb = 1.350
OCZ Ram clock = 1333MhZ
RAM timings = 9-9-9-20-33 and CR 2T
CPU Temps as recorded by Open Hardware Monitor:

CPU - max 54C
Core temp - max 43C
after 2 hours of prime95 - so far no errors

 

[Dave76]

Just checked a FX-4100 review, at full load the CPU should be at ~ 1.368v

Did you increase the CPU Voltage to 1.45v?
I mentioned this previously, just confirming.

What does the CPUZ CPU tab or HWiNFO64 Sensor say for voltage when at idle and under load?

Not any need to raise the CPU voltage any higher than this, you can disable "Cool 'n' Quiet" in your BIOS.
This is what reduces the CPU frequency and voltage when you don't need the full power, when you turn it off the CPU will stay at full frequency and voltage.
You can test this by recreating the crash scenario, do anything that usually causes a crash.

Let us know the Prime95 results.

Download and run OCCT, if it crashes it will save a file with graphs including voltage levels.
If it crashes upload the graphs here.

 

[R4ndom]

Just checked a FX-4100 review, at full load the CPU should be at ~ 1.368v

You mean this review? AMD FX-4100 Quad Core 3.6GHz Bulldozer Processor Review - AMD FX-4100 Turbo Core Details and CPU-Z - Legit Reviews

Did you increase the CPU Voltage to 1.45v?
I mentioned this previously, just confirming.

My MB set it on auto @ 1.440 at stock speed of 3600MhZ it would seem, so quite above the stock voltage. As I thought that was the stock voltage, I added 0.1 as suggested.

Here AMD FX-4100 - CPUs - CPU-Components as in my case, the MB also sets a higher voltage, strangely enough it's a lower version of my MB. Another thread I found AMD FX 4100 BE thread in this case it's a gigabyte board that overvolts the fx-4100.

What does the CPUZ CPU tab or HWiNFO64 Sensor say for voltage when at idle and under load?

The readings are as follows for CPU vcore set @ 1.450 in BIOS:
BIOS HWmonitor - 1.476
HWmonitor64 - minimum 0.964 @ idle and 1.413 @ load
CPU-Z - minumum 1.044 @ idle and 1.476 @ load
Open Hardware Monitor 1.032 @ idle and 1.488 @ load

Not any need to raise the CPU voltage any higher than this, you can disable "Cool 'n' Quiet" in your BIOS.
This is what reduces the CPU frequency and voltage when you don't need the full power, when you turn it off the CPU will stay at full frequency and voltage.
You can test this by recreating the crash scenario, do anything that usually causes a crash.

Let us know the Prime95 results.

Download and run OCCT, if it crashes it will save a file with graphs including voltage levels.
If it crashes upload the graphs here.

Should I then leave CPU voltage @ 1.450 although I saw people getting up to 4.6GhZ with that kind of voltage? Apart from C'n'Q I saw other options in BIOS like C6 state - what should I do with that?

Unfortunately my problem is precisely that I cannot recreate the crash scenario, as it happens randomly, sometimes during idle periods, sometimes just browsing, sometimes just launching an app provokes a crash or BSOD.

For the moment I ran Prime95 with In-place large FFTs test and it didn't hang during something like 2 hours. I'll try the other 2 tests and post the results.

As for OCCT, which of the 2 tests should I run and with which settings? As for voltages reported by OCCT graphs, they have unfortunately the same flaw as HWmonitor64's i.e. 12v is reported to be around 8, which, as you had mentioned would cause that my system wouldn't even boot up. these screenshots are generated every time you stop the app, so I think it's not only when it crashes that you can see graphs - they're saved by default in my documents when exiting the app. Do you want me to upload the screenshots anyway, even if there's no crash?

Thanks!

 

[Dave76]

The different CPU voltages are concerning, just leave it where it's at now.

Let's disable C 'n Q for now, this controls CPU frequency and voltage depending on use. Which seems to be more likely considering the description of the crashes you're having. With this disabled your CPU will run at max frequency and voltage at all times, this will tell us if there are issues when the CPU frequency/voltages change.
The C6 states are sleep/hibernate settings, may check them later.

Your crash scenario seems to indicate that it might be when the CPU frequency/voltage change it causes a crash, so test by disabling C 'n Q.

Run OCCT - Large Data Set, not LINPACK.
Even though it may be giving false voltage readings, watch the Voltage graph for changes in the CPU, +12V, +5V and +3.3V, readings. You are looking for drops and fluctuations, these may be causing the crashes.

 

[R4ndom]

Hi,

After some testing (still with C'n'Q enabled) and other options left as described in my previous post, I didn't have a single BSOD or app crash for 2-3 days. This however only lasted until this morning, where I had a lovely BSOD again - dump attached, without any apparent reason - I was browsing the net, that's all.

The event viewer reports the following:

Event name: APPCRASH
Reply: unavailable
Cab file ID: 0

Problem's signature:
P1: svchost.exe_SysMain
P2: 6.1.7600.16385
P3: 4a5bc3c1
P4: sysmain.dll
P5: 6.1.7601.17514
P6: 4ce7c9db
P7: c0000005
P8: 000000000003d9dc
P9: 
P10: 

Attached files: 
C:\Windows\Temp\WERFFB2.tmp.appcompat.txt
C:\Windows\Temp\WERFFE1.tmp.WERInternalMetadata.xml
C:\Windows\Temp\WERFFE2.tmp.hdmp
C:\Windows\Temp\WER1E6.tmp.mdmp

These files may be available here:
C:\ProgramData\Microsoft\Windows\WER\ReportQueue\AppCrash_svchost.exe_SysM_91bbaabfdaa7ebd4bf6f962c64299c3c345759_cab_0f9a02cd

Analysis symbol: 
Solution re-check: 0
Report ID: a0816db0-a7da-11e1-9bd6-14dae9c5e52a

Prime95 passed all the 3 tests - all in all 18 hours of continuous testing. Same for OCCT - passed both tests which ran for around 10 hours each (separately from Prime95 - i.e. not at the same time).

Today I disabled the C'n'Q and I had to disable the C6 states as well, otherwise it seems the voltages would change according to the load.

I attach the OCCT screenshots - can you observe anything unusal? this is still from when the C'n'Q and C6 were enabled.

I'll run today OCCT with C'nQ disabled, although again, crashes never happen with PC under load.

thanks!

 

[zigzag3143]

Just in passing.

You have a driver installed that is known to cause memory management crashes called asacpi.sys. Yours is from 2005


Asacpi.sys

The pre 2009 version of this driver is a known BSOD cause.

Please visit this link: Asus tek computer inc. -support- drivers and download p7p55d le

ASUSTeK Computer Inc. -Support- Drivers and Download P7P55D LE
ASUSTeK Computer Inc. - Motherboards- ASUS P5K-VM

Scroll down to the utilities category, then scroll down to the "atk0110 driver for windowsxp/vista/windows 7 32&64-bit" (it's about the 12th item down).

Download and install it.

Go to c:\windows\system32\drivers to check and make sure that the asacpi.sys file is date stamped from 2009 or 2010 (not before).

 

[R4ndom]

Thanks, I updated the driver and the current version is from May 2009.

I start again OCCT 4.2.0 version with large data set option.

 

[R4ndom]

The OCCT finished without errors and ran for 6 hours, I attach the screenshots.

This was done with C'n'Q and C6 disabled as well as all hibernation/sleep options disabled in windows.

I set the voltage a bit lower in BIOS before launching the test - 1.3875.

 

[Dave76]

The first OCCT files shows low voltage spikes on the CPU VCore graph, seems to be random.
The voltage should be steady with some fluctuation, there seems to be a lot of voltage variation.
My OCCT CPU VCore graphs are flat single lines, yours are showing fluctuations from 1.37 to 1.48 with low spike to 1.25.
Your second CPU VCore is showing fluctuations from 1.315 to 1.41.
This is telling me the power is not stable.

Let's confirm with a shorter test, so we can see the fluctuations better.
Run two more OCCT tests, one with C'nQ enabled (same settings as the first previous test), then the second test with C'nQ diabled (same settings as the second previous test).

Leave it at idle for 5 minutes, then run the same test as before, but for 1 hour only.
I want to get a better look at the CPU VCore voltages.

Here's what mine looks like:
View attachment 213934

Yours:
View attachment 213935

 

[R4ndom]

Hi,
I haven't had time to do the test you suggested, I'll post the results as soon as they're done.

Today however I saw a strange thing. When I logged into windows, I checked the event viewer (became almost a habit now for me) I saw more than a 100 entries marked as "windows error reporting" and "appcrash" "apphang" and "bex", however their "level" was indicated as "information" and not "error" with an exclamation mark. But the thing is that NONE of these apps was launched at the time windows reported errors! At least not all of them, as I don't know which windows services were running at logon. The list is long, but I'll give a few examples.

svchost.exe_SysMain, sysmain.dll - these two at least 20 times within 1-2sec intervals, then iw3mp.exe - this was not launched at all, explorer.exe (apphang) although I didn't experience any hang, SopCast.exe - wasn't launched, PRIME95.EXE - wasn't running, i_view32.exe - wasnt' running, firefox.exe - wasn't running, WinFlash.exe, Skype.exe - wasn't running, pidgin.exe - wasn't running, Mahjong.exe - never launched it since I installed windows, java.exe, iexplore.exe, Explorer.EXE, Dwm.exe, Microsoft Windows.NetworkDiagnostics.1.0, and so on. So now it's getting even funnier, as things crash even before they're launched

So I tried sfc, it said it couldn't repair certain files. I booted from WinDVD to scan it with offboot/offwindir command. It said the same, it couldn't repair certain files. I looked at the CBS.log file and here's what I found:

2012-05-29 17:32:11, Info                  CSI    000002ee [SR] Verify complete
2012-05-29 17:32:11, Info                  CSI    000002ef [SR] Repairing 6 components
2012-05-29 17:32:11, Info                  CSI    000002f0 [SR] Beginning Verify and Repair transaction
2012-05-29 17:32:11, Info                  CSI    000002f1 [SR] Cannot repair member file [l:76{38}]"Rights-Management-Client-v1-API-DL.man" of Microsoft-Windows-Migration-DownlevelManifests, Version = 6.1.7601.17514, pA = PROCESSOR_ARCHITECTURE_AMD64 (9), Culture neutral, VersionScope = 1 nonSxS, PublicKeyToken = {l:8 b:31bf3856ad364e35}, Type neutral, TypeName neutral, PublicKey neutral in the store, file is missing
2012-05-29 17:32:11, Info                  CSI    000002f2 [SR] Cannot repair member file [l:50{25}]"VaultCredProvider.dll.mui" of Microsoft-Windows-Security-Vault.Resources, Version = 6.1.7600.16385, pA = PROCESSOR_ARCHITECTURE_AMD64 (9), Culture = [l:10{5}]"pl-PL", VersionScope = 1 nonSxS, PublicKeyToken = {l:8 b:31bf3856ad364e35}, Type neutral, TypeName neutral, PublicKey neutral in the store, file is missing
2012-05-29 17:32:11, Info                  CSI    000002f3 [SR] Cannot repair member file [l:36{18}]"Amd64\BRM7840N.GPD" of prnbr006.inf, Version = 6.1.7600.16385, pA = PROCESSOR_ARCHITECTURE_AMD64 (9), Culture neutral, VersionScope = 1 nonSxS, PublicKeyToken = {l:8 b:31bf3856ad364e35}, Type = [l:24{12}]"driverUpdate", TypeName neutral, PublicKey neutral in the store, file is missing
2012-05-29 17:32:11, Info                  CSI    000002f4 [SR] Cannot repair member file [l:36{18}]"Amd64\hpc4600t.vdf" of prnhp005.inf, Version = 6.1.7600.16385, pA = PROCESSOR_ARCHITECTURE_AMD64 (9), Culture neutral, VersionScope = 1 nonSxS, PublicKeyToken = {l:8 b:31bf3856ad364e35}, Type = [l:24{12}]"driverUpdate", TypeName neutral, PublicKey neutral in the store, file is missing
2012-05-29 17:32:11, Info                  CSI    000002f5 [SR] Cannot repair member file [l:20{10}]"aeevts.dll" of Microsoft-Windows-Application-Experience-Common-Logger, Version = 6.1.7600.16385, pA = PROCESSOR_ARCHITECTURE_INTEL (0), Culture neutral, VersionScope = 1 nonSxS, PublicKeyToken = {l:8 b:31bf3856ad364e35}, Type neutral, TypeName neutral, PublicKey neutral in the store, file is missing
2012-05-29 17:32:12, Info                  CSI    000002f6 [SR] Cannot repair member file [l:16{8}]"WMIC.exe" of Microsoft-Windows-WMI-Command-Line-Utility, Version = 6.1.7600.16385, pA = PROCESSOR_ARCHITECTURE_INTEL (0), Culture neutral, VersionScope = 1 nonSxS, PublicKeyToken = {l:8 b:31bf3856ad364e35}, Type neutral, TypeName neutral, PublicKey neutral in the store, file is missing
2012-05-29 17:32:12, Info                  CSI    000002f7 [SR] Cannot repair member file [l:36{18}]"Amd64\BRM7840N.GPD" of prnbr006.inf, Version = 6.1.7600.16385, pA = PROCESSOR_ARCHITECTURE_AMD64 (9), Culture neutral, VersionScope = 1 nonSxS, PublicKeyToken = {l:8 b:31bf3856ad364e35}, Type = [l:24{12}]"driverUpdate", TypeName neutral, PublicKey neutral in the store, file is missing
2012-05-29 17:32:12, Info                  CSI    000002f8 [SR] This component was referenced by [l:186{93}]"Microsoft-Windows-Printer-Drivers-Package~31bf3856ad364e35~amd64~~6.1.7601.17514.INF_prnbr006"
2012-05-29 17:32:12, Info                  CSI    000002f9 [SR] Cannot repair member file [l:36{18}]"Amd64\hpc4600t.vdf" of prnhp005.inf, Version = 6.1.7600.16385, pA = PROCESSOR_ARCHITECTURE_AMD64 (9), Culture neutral, VersionScope = 1 nonSxS, PublicKeyToken = {l:8 b:31bf3856ad364e35}, Type = [l:24{12}]"driverUpdate", TypeName neutral, PublicKey neutral in the store, file is missing
2012-05-29 17:32:12, Info                  CSI    000002fa [SR] This component was referenced by [l:186{93}]"Microsoft-Windows-Printer-Drivers-Package~31bf3856ad364e35~amd64~~6.1.7601.17514.INF_prnhp005"
2012-05-29 17:32:12, Info                  CSI    000002fb [SR] Cannot repair member file [l:20{10}]"aeevts.dll" of Microsoft-Windows-Application-Experience-Common-Logger, Version = 6.1.7600.16385, pA = PROCESSOR_ARCHITECTURE_INTEL (0), Culture neutral, VersionScope = 1 nonSxS, PublicKeyToken = {l:8 b:31bf3856ad364e35}, Type neutral, TypeName neutral, PublicKey neutral in the store, file is missing
2012-05-29 17:32:12, Info                  CSI    000002fc [SR] This component was referenced by [l:202{101}]"Microsoft-Windows-Foundation-Package~31bf3856ad364e35~amd64~~6.1.7601.17514.WindowsFoundationDelivery"
2012-05-29 17:32:12, Info                  CSI    000002fd [SR] Cannot repair member file [l:76{38}]"Rights-Management-Client-v1-API-DL.man" of Microsoft-Windows-Migration-DownlevelManifests, Version = 6.1.7601.17514, pA = PROCESSOR_ARCHITECTURE_AMD64 (9), Culture neutral, VersionScope = 1 nonSxS, PublicKeyToken = {l:8 b:31bf3856ad364e35}, Type neutral, TypeName neutral, PublicKey neutral in the store, file is missing
2012-05-29 17:32:12, Info                  CSI    000002fe [SR] This component was referenced by [l:242{121}]"Microsoft-Windows-Client-Features-Package~31bf3856ad364e35~amd64~~6.1.7601.17514.Microsoft-Windows-Client-Features-Update"
2012-05-29 17:32:13, Info                  CSI    000002ff [SR] Cannot repair member file [l:50{25}]"VaultCredProvider.dll.mui" of Microsoft-Windows-Security-Vault.Resources, Version = 6.1.7600.16385, pA = PROCESSOR_ARCHITECTURE_AMD64 (9), Culture = [l:10{5}]"pl-PL", VersionScope = 1 nonSxS, PublicKeyToken = {l:8 b:31bf3856ad364e35}, Type neutral, TypeName neutral, PublicKey neutral in the store, file is missing
2012-05-29 17:32:13, Info                  CSI    00000300 [SR] This component was referenced by [l:266{133}]"Microsoft-Windows-WindowsFoundation-LanguagePack-Package~31bf3856ad364e35~amd64~pl-PL~6.1.7601.17514.Windows Foundation Language Pack"
2012-05-29 17:32:13, Info                  CSI    00000301 [SR] Cannot repair member file [l:16{8}]"WMIC.exe" of Microsoft-Windows-WMI-Command-Line-Utility, Version = 6.1.7600.16385, pA = PROCESSOR_ARCHITECTURE_INTEL (0), Culture neutral, VersionScope = 1 nonSxS, PublicKeyToken = {l:8 b:31bf3856ad364e35}, Type neutral, TypeName neutral, PublicKey neutral in the store, file is missing
2012-05-29 17:32:13, Info                  CSI    00000302 [SR] This component was referenced by [l:202{101}]"Microsoft-Windows-Foundation-Package~31bf3856ad364e35~amd64~~6.1.7601.17514.WindowsFoundationDelivery"
2012-05-29 17:32:13, Info                  CSI    00000303 Repair results created:

Can you make anything out of it? I googled for some names, but apart from WMIC.exe I couldn't find much info... explorer.exe file is listed many times and at least with 2 different spellings - is it normal?

edit: I found this How to use the System File Checker tool to troubleshoot missing or corrupted system files on Windows Vista or on Windows 7 should I try it? could it be that my Win install DVD is corrupt and so are the files, which in consequence causes all the mess I've been experiencing with my pc? if so, is it possible to check the integrity of the DVD to confirm or rule out that it's the culprit?

 

[Dave76]

The testing and results are leading me to a PSU problem.
The OCCT CPU VCore test should confirm this.

You can run this tutorial, option three: How to Repair Windows 7 System Files with System File Checker

Run File Hash Checker on your Win7 ISO: File Hash Checker 1.1

 

[R4ndom]

Hi Dave and thanks for answering. I have a couple of questions though:
How to check with the programme you sent me the link to if my Win DVD isn't corrupted? It only offers to check individual files and not the whole DVD. I made an ISO out of it with WinISO and ran a check on this. However, I can't find any reference to MD5/SHA1 sums for my version of Windows.

As for repairing according to MS tutorial, the files listed as corrupted and unrepairable by sfc, they reside in many different folders within C:\windows for example aeevts.dll is present in 4 different folders and I have no clue which has to be replaced and I have no clue either where to get uncorrupted versions from...

I attach results from today with C'n'Q, turbo boost and C6 ON and OFF in 2 separate folders inside the zip.

Can the voltage fluctuations be due to MB instability as well PSU's and are they above normal? vcore in BIOS is set to 1.3875v

 

[Dave76]

List of known MD5/SHA1 for windows OS

Windows 7 Home Premium 64-bit:
Name: en_windows_7_home_premium_x64_dvd_x15-65733.iso
SHA-1: 336779EA6B65F63E11A609B4D021439C47AB315B
CRC: 56D954E4

If you made the ISO from a DVD it may not be the same.
You can D/L the Windows 7 home Premium 64bit ISO from MS and use your key.
Easy way is to put it on a USB flash drive.
Windows 7 USB/DVD Download Tool

To repair files that are not fixed by SFC/ scannow it is best to do a repair install, that will get all the corrupt files.
How to Do a Repair Install to Fix Windows 7

The voltage fluctuations can be due to MB instability, but more likely caused by the PSU.
C'nQ, C6 and Turbo on:

• +3.3v is not stable
• +5v is not stable
• +12v is not stable
• CPU Vcore is horrible

The CPU Vcore was very unstable even when the test was over.

C'nQ, C6 and Turbo off:
Same, just as bad as when the settings were on, this proves it is not likely the motherboard.
The CPU Vcore voltage is very bad, it's remarkable that the system is running at all.

Replace the PSU, get a good brand with headroom for any upgrades you may do in the future.
A bad PSU can permanently damage all the other components, CPU, RAM, Graphics card, motherboard, Monitor... everything.

If this was my computer, I wouldn't start it again until the new PSU was installed.

 

[R4ndom]

Hi,

thanks for the links to the repair install, I will try this for sure at some point and will let you know the result. I cannot find the right sums for my version of Windows, so I think I'll download one from MS site, I'm just wondering if OEM key will allow me do that.

As for voltages, I read somewhere that it's actually the MB that manages the CPU voltage (can you confirm that?) and while the PSU fail will surely lead to problems, the CPU vcore fluctiations are due to MB. On mine it seems that it could possibly not only supply unstable voltage, but it's sensors are also worth close to nothing.

I checked all the voltages, but this time not basing myself on buggy software reading from buggy sensors and resulting in 8V for the 12V line as seen on the OCCT graph.

I performed checks on the three lines 12V, 5V and 3.3V under stress and on idle using a multimeter which I plugged directly to the molexes.

Each reading was done for approximately 15-20s for each line and the results were as follows:
stress:

12V: 12.10V - 12.13V
5V: 5.09V - 5.12V
3.3V: 3.35V

idle:

12V: 12.11V-12.15V
5V: 5.10V - 5.13V
3.3V: 3.36V

As you see fluctuations are minimal, if not non-existent. Anyway, they're more than correct according to the ATX norm which allows for 5% variations for 12V and 5V lines and 4% for the 3.3V line. So could it point to the conclusion that PSU isn't the problem in my case after all?..

 

[R4ndom]

I decided to start from scratch and reinstalled windows. I first tried the repair install, but it wouldn't work saying that it was missing some files and would freeze after some time.

So I reformatted the drive and did a clean install.

Already there, even before I installed chipset drivers, gpu drivers and all the others, as well as antivir (MSE), I got a BSOD just like this. I was actually in event viewer to see WMI errors and some other I can't recall now.

So I knew already at that point that it's useless to continue, nevertheless I installed the required drivers, then tried to update windows, it would hang upon installation of some updates, around 14 out of 51. I waited for 2 or 3 hours, and then had to do a hard reset.

In the meantime, internet explorer would crash, as well as windows explorer.

I restarted windows update, it hang again and I got 2 more bsods in a couple of minutes.

Maybe the chronology isn't perfect here, but it's enough I think to see that something is definitely wrong with this pc. WhoCrashed! reported "a third party driver that might have cause the 1st BSOD" but I didn't have ANY other drivers installed at that time, there was only Windows installed.

I will attach BSODs dumps this afternoon.

Another weird thing: after every blue screen, upon next login, you should normally get a window saying: windows has recovered after an unexpected shutdown". I don't see the message and NOTHING is recorded in the event viewer concerning the BSOD. On the other hand, there are loads of errors relating to appcrashes - mainly svchost, iexplorer, windows explorer.

I cannot download the windows 7 home premium iso from the link you provided, seems like MS won't let you download an iso unless you buy an electronically distributed version of windows.... and this can be done only for specific language versions, mine has never been on sale through electronic channel.

 

[R4ndom]

Here's the dump file. Last BSOd took place when I was running the jcgriff script and after this BSOD only I received a message when logging in again to windows that "it has recovered after an unexpected shutdown". There's also only 1 single entry in event viewer relating to BSODs since I did the fresh install 2 days ago and thus its the only entry in the jcgriff zip. All of the other BSODS weren't recorded at all. I borrowed a disk today to try yet another install to rule out disk issues.

Strangely, the WhoCrashed! app detected 2 different BSODS:

[FONT=Segoe UI]Crash dump directory: C:\Windows\Minidump

[/FONT][SIZE=2][COLOR=#000000][FONT=Segoe UI][SIZE=2][COLOR=#000000]Crash dumps are enabled on your computer.
[/COLOR][/SIZE][/FONT][/COLOR][/SIZE]

[FONT=Segoe UI,  Arial][SIZE=2][COLOR=#000000][B]On Mon 2012-06-18 16:06:43 GMT your computer crashed[/B]
crash dump file: C:\Windows\Minidump\061812-21684-01.dmp
This was probably caused by the following module: [FONT=Segoe UI,  Arial][SIZE=2][COLOR=#0066cc][URL="http://www.google.com/search?q=ntoskrnl.exe"][U]ntoskrnl.exe[/U][/URL][/COLOR][/SIZE][/FONT] (nt+0x1AB007) 
Bugcheck code: 0x1000007E (0xFFFFFFFFC0000005, 0xFFFFF80002E08007, 0xFFFFF880031B6778, 0xFFFFF880031B5FD0)
Error: [FONT=Segoe UI,  Arial][SIZE=2][COLOR=#0066cc][URL="http://www.google.com/search?q=MSDN+bugcheck+SYSTEM_THREAD_EXCEPTION_NOT_HANDLED_M"][U]SYSTEM_THREAD_EXCEPTION_NOT_HANDLED_M[/U][/URL][/COLOR][/SIZE][/FONT]
file path: C:\Windows\system32\ntoskrnl.exe
product: [FONT=Segoe UI,  Arial][SIZE=2][COLOR=#0066cc][URL="http://www.google.com/search?q=Microsoft® Windows® Operating System"][U]Microsoft® Windows® Operating System[/U][/URL][/COLOR][/SIZE][/FONT]
company: [FONT=Segoe UI,  Arial][SIZE=2][COLOR=#0066cc][URL="http://www.google.com/search?q=Microsoft Corporation"][U]Microsoft Corporation[/U][/URL][/COLOR][/SIZE][/FONT]
description: NT Kernel & System
Bug check description: This indicates that a system thread generated an exception which the error handler did not catch.
This appears to be a typical software driver bug and is not likely to be caused by a hardware problem. 
The crash took place in the Windows kernel. Possibly this problem is caused by another driver which cannot be identified at this time. 
[/COLOR][/SIZE][/FONT]

[FONT=Segoe UI,  Arial][SIZE=2][COLOR=#000000][B]On Mon 2012-06-18 16:06:43 GMT your computer crashed[/B]
crash dump file: C:\Windows\memory.dmp
This was probably caused by the following module: [FONT=Segoe UI,  Arial][SIZE=2][COLOR=#0066cc][URL="http://www.google.com/search?q=ntkrnlmp.exe"][U]ntkrnlmp.exe[/U][/URL][/COLOR][/SIZE][/FONT] (nt!KeBugCheckEx+0x0) 
Bugcheck code: 0x7E (0xFFFFFFFFC0000005, 0xFFFFF80002E08007, 0xFFFFF880031B6778, 0xFFFFF880031B5FD0)
Error: [FONT=Segoe UI,  Arial][SIZE=2][COLOR=#0066cc][URL="http://www.google.com/search?q=MSDN+bugcheck+SYSTEM_THREAD_EXCEPTION_NOT_HANDLED"][U]SYSTEM_THREAD_EXCEPTION_NOT_HANDLED[/U][/URL][/COLOR][/SIZE][/FONT]
Bug check description: This bug check indicates that a system thread generated an exception that the error handler did not catch.
The crash took place in the Windows kernel. Possibly this problem is caused by another driver which cannot be identified at this time. 
[/COLOR][/SIZE][/FONT]

 

[R4ndom]

As I was trying to reply to the thread, I had another BSOD and before that one, yet another as I tried to open Windows Management Console. WhoCrashed! reported the following:

[FONT=Segoe UI][B]On Mon 2012-06-18 16:40:53 GMT your computer crashed[/B][/FONT]
[FONT=Segoe UI]crash dump file: C:\Windows\Minidump\061812-19375-01.dmp[/FONT]
[FONT=Segoe UI]This was probably caused by the following module: [SIZE=2][COLOR=#0066cc][URL="http://www.google.com/search?q=ntoskrnl.exe"][U]ntoskrnl.exe[/U][/URL][/COLOR][/SIZE] (nt+0x7F1C0) [/FONT]
[FONT=Segoe UI]Bugcheck code: 0x4E (0x99, 0x120DF6, 0x2, 0x1201F5)[/FONT]
[FONT=Segoe UI]Error: [SIZE=2][COLOR=#0066cc][URL="http://www.google.com/search?q=MSDN+bugcheck+PFN_LIST_CORRUPT"][U]PFN_LIST_CORRUPT[/U][/URL][/COLOR][/SIZE][/FONT]
[FONT=Segoe UI]file path: C:\Windows\system32\ntoskrnl.exe[/FONT]
[FONT=Segoe UI]product: [SIZE=2][COLOR=#0066cc][URL="http://www.google.com/search?q=Microsoft® Windows® Operating System"][U]Microsoft® Windows® Operating System[/U][/URL][/COLOR][/SIZE][/FONT]
[FONT=Segoe UI]company: [SIZE=2][COLOR=#0066cc][URL="http://www.google.com/search?q=Microsoft Corporation"][U]Microsoft Corporation[/U][/URL][/COLOR][/SIZE][/FONT]
[FONT=Segoe UI]description: NT Kernel & System[/FONT]
[FONT=Segoe UI]Bug check description: This indicates that the page frame number (PFN) list is corrupted.[/FONT]
[FONT=Segoe UI]This appears to be a typical software driver bug and is not likely to be caused by a hardware problem. This might be a case of memory corruption. More often memory corruption happens because of software errors in buggy drivers, not because of faulty RAM modules. [/FONT]
[FONT=Segoe UI]The crash took place in the Windows kernel. Possibly this problem is caused by another driver which cannot be identified at this time. [/FONT]
 
 
[SIZE=2][COLOR=#000000][FONT=Segoe UI][B]On Mon 2012-06-18 16:40:53 GMT your computer crashed[/B][/FONT][/COLOR][/SIZE]
[SIZE=2][FONT=Segoe UI][COLOR=#000000]crash dump file: C:\Windows\memory.dmp[/COLOR][/FONT][/SIZE]
[SIZE=2][FONT=Segoe UI][COLOR=#000000]This was probably caused by the following module: [SIZE=2][COLOR=#0066cc][URL="http://www.google.com/search?q=ntkrnlmp.exe"][U]ntkrnlmp.exe[/U][/URL][/COLOR][/SIZE] (nt!KeBugCheckEx+0x0) [/COLOR][/FONT][COLOR=#000000]
[FONT=Segoe UI]Bugcheck code: 0x4E (0x99, 0x120DF6, 0x2, 0x1201F5)[/FONT]
[FONT=Segoe UI]Error: [SIZE=2][COLOR=#0066cc][URL="http://www.google.com/search?q=MSDN+bugcheck+PFN_LIST_CORRUPT"][U]PFN_LIST_CORRUPT[/U][/URL][/COLOR][/SIZE][/FONT]
[FONT=Segoe UI]Bug check description: This indicates that the page frame number (PFN) list is corrupted.[/FONT]
[FONT=Segoe UI]This appears to be a typical software driver bug and is not likely to be caused by a hardware problem. This might be a case of memory corruption. More often memory corruption happens because of software errors in buggy drivers, not because of faulty RAM modules. [/FONT]
[FONT=Segoe UI]The crash took place in the Windows kernel. Possibly this problem is caused by another driver which cannot be identified at this time. [/FONT]
[/COLOR][/SIZE]
 
[FONT=Segoe UI,  Arial][SIZE=2][COLOR=#000000][B]On Mon 2012-06-18 16:32:52 GMT your computer crashed[/B][/COLOR][/SIZE][/FONT]
[SIZE=2][FONT=Segoe UI,  Arial][COLOR=#000000]crash dump file: C:\Windows\Minidump\061812-21855-01.dmp[/COLOR][/FONT][/SIZE]
[SIZE=2][FONT=Segoe UI,  Arial][COLOR=#000000]This was probably caused by the following module: [FONT=Segoe UI,  Arial][SIZE=2][COLOR=#0066cc][URL="http://www.google.com/search?q=ntoskrnl.exe"][U]ntoskrnl.exe[/U][/URL][/COLOR][/SIZE][/FONT] (nt+0x7F1C0) 
Bugcheck code: 0x19 (0x20, 0xFFFFF8A008E56680, 0xFFFFF8A008E56CD0, 0x5650074)
Error: [FONT=Segoe UI,  Arial][SIZE=2][COLOR=#0066cc][URL="http://www.google.com/search?q=MSDN+bugcheck+BAD_POOL_HEADER"][U]BAD_POOL_HEADER[/U][/URL][/COLOR][/SIZE][/FONT]
file path: C:\Windows\system32\ntoskrnl.exe
product: [FONT=Segoe UI,  Arial][SIZE=2][COLOR=#0066cc][URL="http://www.google.com/search?q=Microsoft® Windows® Operating System"][U]Microsoft® Windows® Operating System[/U][/URL][/COLOR][/SIZE][/FONT]
company: [FONT=Segoe UI,  Arial][SIZE=2][COLOR=#0066cc][URL="http://www.google.com/search?q=Microsoft Corporation"][U]Microsoft Corporation[/U][/URL][/COLOR][/SIZE][/FONT]
description: NT Kernel & System
Bug check description: This indicates that a pool header is corrupt.
This appears to be a typical software driver bug and is not likely to be caused by a hardware problem. This might be a case of memory corruption. More often memory corruption happens because of software errors in buggy drivers, not because of faulty RAM modules. 
The crash took place in the Windows kernel. Possibly this problem is caused by another driver which cannot be identified at this time. 
[/COLOR][/FONT][/SIZE]

Hope you can read it before it bsods again. It's even worse than before OS reinstall. I begin installing on the borrowed disk.

 

[writhziden]

All your crashes indicate the RAM as the culprit still. Sorry about my original response. Seems I found the wrong RAM. I now see the link you gave: OCZ DDR3 PC3-12800 CL8 Edition Reaper Low Voltage Dual Channel - OCZ


Dave highly suspects the PSU. His expertise with hardware is difficult to match, so I would be pretty confident that he knows what he is talking about. If you do not want to replace the PSU, at least try to find a PSU you can borrow that has similar wattage or better (preferably better) to your current PSU to test. You should try to borrow from a friend or colleague who is willing to lend you the PSU for quick testing purposes.


Also, since I am somewhat new to this thread, can you list what hardware you have replaced since starting the thread?

 

[R4ndom]

All your crashes indicate the RAM as the culprit still. Sorry about my original response. Seems I found the wrong RAM. I now see the link you gave: OCZ DDR3 PC3-12800 CL8 Edition Reaper Low Voltage Dual Channel - OCZ

Hi writhziden and thanks for the reply. Yes, that's the RAM I have, what do you mean by your original response? The one before edit? If so, I didn't see it, so don't worry

Dave highly suspects the PSU. His expertise with hardware is difficult to match, so I would be pretty confident that he knows what he is talking about. If you do not want to replace the PSU, at least try to find a PSU you can borrow that has similar wattage or better (preferably better) to your current PSU to test. You should try to borrow from a friend or colleague who is willing to lend you the PSU for quick testing purposes.

Me too was suspecting the PSU, especially after what has been concluded concerning voltages, etc. But then I checked the PSU with a multimeter and showed no fluctuations whatsoever - 0.1-03V are marginal to the extent that they could be due to measurement errors in the device itself.

We were suspecting RAMs, so I bought new ones, which appeared to be somewhat odd, as my PC would keep crashing when I installed them and would generate errors in memtest after 2-3 minutes and the next day it would pass memtest for 17 hours straigth without a single error. So I came back to using the OCZ and I think I won't even send the others for RMA as it would cost me the same price I paid just to post the sticks.

MB was suspected as well, but I think that's the element most difficult to rule out, most of the time it either works or not.

CPU had also been a suspect, but I re-seated it twice since january, no bent pins, temps 32 idle and 52max stress, could always be faulty electronics, but Prime tests showed no errors.

The disk has never been considered as the failing component. However, yesterday I could install win7 on the borrowed disk without any hiccups and here's what I concluded:

1) It installed quicker than on the old disk I would say.
2) All the windows updates installed seamlessly, while with the other disk it was a continuous blue screen, reinstall of updates, freeze or crash of windows updates module and finally another blue screen, all this without any visible pattern and order
3) UAC was failing - this I can see now with the fresh install, some functions just didn't exist with the old install. i.e. it asked me for permissions when I tried to rename a folder, this wasn't the cae with previous installations
4) I get no more svchost app crashes in event viewer, explorer and iexplorer crashes/hangs, windows management console errors are gone as well
5) MSE now starts every time, and so does LAN connection - this wasn't the case before, when MSE would report it's off, and LAN connection desactivated.

All this is weird, as chkdsk and seatools never detected any problems, ok, occasionnaly it would repair something, but SMART readings, surface tests, etc. all showed no signs of corruption. How is it possible then that still the drive is corrupted in some way?

I still want to test it for some days before returning the working disk to the friend, but all signs on earth and heaven point to the samsung as the culprit.

so, since I am somewhat new to this thread, can you list what hardware you have replaced since starting the thread?

Well, since the start of the thread, I replaced RAMs, re-seated all other components, and yesterday borrowed the disk on which I installed the latest OS. Fortunately I didn't buy a new PSU, and decided to test with another hdsk This was in fact already suggested by Dave or you at some point, but I didn't have access to any hd I could borrow at the time. It's a pity, because probably the whole thread would be half of what it's now or less. Presuming of course that it's really the drive that's at fault. I'll keep testing to see, I will also add programmes to see if corruption doesn't reappear. Thanks again for all your contributions.

 

[writhziden]

Nice work with the hard disk swap!

I too wonder why the diagnostics did not reveal anything if it does end up being the hard disk is the culprit. The only thing I can think of is that the hard drive may have just begun to fail, so diagnostics were picking up on the fact that any errors they found were within the realm of normal operating use. It is normal for bad sectors to occur from time to time on any hard disk; in fact, hard disks are designed with spare sectors for this reason. It is when bad sectors start accumulating over time that one should start to worry.

Also, if corruption is occurring on a regular basis, that is another sign of a failing hard disk and may not be picked up by diagnostics since corruption is also within the realm of normal hard disk use. I am not sure what would cause such behavior, but a possible cause would be a slightly bad hard disk head that only causes issues every once in a while but not during extended diagnostic tests. That scenario is unlikely, but possible. The more likely cause is an inconsistent platter or spindle that throws off the head as it writes so it corrupts a sector (or more) during writing.

Diagnostic utilities are not foolproof, unfortunately. They have to take into account a lot of factors that are within operating standards, so if one of those factors is occurring more than it should but is still within specs, the diagnostics may miss it.

Further reading: NTFS.com HARD DRIVES. Hard Disk Drive Basics.