New MoBo to replace the former, with LGA1155 socket, because of BSODs

[Da Rossa]

I've been experiencing blue screen problems for about six months now. They're more frequent when I'm playing Starcraft 2 with voice chat. Actually, they only happen when I'm at SC, but Blizzard guys claim this problem is outside their scope. They told me to check some Microsoft KB articles, which are inconveniently vague: each BSOD problem can have at least four distinct causes.

The errors I've been getting are PAGE_FAULT_IN_NONPAGED_AREA and KERNEL_SECURITY_CHECK_FAILURE, and a third one I don't recall right now. No one knows their cause in my system. And it is difficult to reproduce the problem in order to insulate it. It's like a cancer, which affects any person at any age. To make things even more complicated, there are no Windows dump files were they were supposed to be (C:\windows\minidump).

After investigating, I found out the errors could be due to the audio controller on the motherboard. In fact, I already tried to update the drivers, but this persists:

Meaning on the right: "This device could not be initiated/started (Code 10). Operation was not succeeded."

Maybe it is a case in which I should replace my motherboard.

My Setup:
Processor: Intel Core i5 3450 @3.1GHz
MoBo: Gigabyte Z77M-D3H
Memory: 8192MB DDR3 Corsair 1333MHz
VGA: Saphire ATI Radeon HD 7770 1024MB 128 bits @1920x1080
SSD: Kingston SUV400S372240G 240GB
Hard Drive 2: Seagate ST3500413AS 465 GB 7200RPM
Hard Drive 3: Seagate ST3750330AS 698 GB 7200RPM
Hard Drive 4: Western Digital WD20EARX 1863GB IntelliPower
Hard Drive 5: Western Digital WD20EZRX 1863GB IntelliPower
Power Supply: Corsair VX450
OS: Windows 10 Home
Monitor: Samsung T270B350 @1920x1080

The "core" of the above system was bought in 2012. The motherboard has a LGA1155 socket, which seems to have been discontinued. In order to avoid buying new processor and memory sticks, I'm thinking about buying just a new motherboard, but preferably from another brand (I like Asus); a board with a good built-in sound controller (are there still offboard sound cards on the market?) So I ask:

1. Should I really proceed to buying a new motherboard or should I investigate even further the cause for the dreaded BSOD?
2. What model to get, having in mind the age of the socket?

Thanks!

 

[captaincharisma]

its going to be difficult to find a motherboard for a 3rd gen i5 CPU like that. beleve me i tried but all i could find was low quality Chinese knockoff mobos on ebay and gave up and just waited till i was able to afford to get a new CPU/mobo. but if it is just the sound chip why not try getting a dedicated sound card from creative labs or asus.

 

[Da Rossa]

I don't know, the audio controller is a suspect, but not confirmed. If I acquire a new gear just as an attempt to isolate the problem, and it persists, it will be a waste. Buying a new motherboard will be a waste too except that I see no other solution

 

[captaincharisma]

I don't know, the audio controller is a suspect, but not confirmed. If I acquire a new gear just as an attempt to isolate the problem, and it persists, it will be a waste. Buying a new motherboard will be a waste too except that I see no other solution

tru, but a sound card would be cheaper then a motherboard

 

[Da Rossa]

Apparently not here in Brazil. Today very few people use dedicated sound cards, so the market has realised it's now a "niche" product, hence, too expensive for its secondary role.

 

[Corwin65]

Look for USB soundcards. They plug into the USB port and give you an out and a mic. Can get them here in the states for as little as $5.

Here is an example - https://www.amazon.com/Virtual-Channel-Audio-Adapter-Notebook/dp/B00M3UWE3Q/ref=sr_1_14?s=electronics&ie=UTF8&qid=1513791862&sr=1-14&keywords=usb+sound+card

 

[Da Rossa]

Ok you convinced me not to go for a new motherboard. So do you have any clues regarding the blue screen errors? Or should I open another thread for that?

 

[Corwin65]

Right click it in device manager and see if you you can disable it.

 

[johnbl]

if your system:
is set to save a minidump, has a pagefile.sys, and has proper free space on the drive it should create a memory dump file in c:\windows\minidump directory.
if it is set to save a kernel dump it will create a file c:\windows\memory.dmp

if it does not create a memory dump, then something in the storage system is mess up.
(update the drivers from the motherboard vendors website)
then google "how to force a memory dump using a keyboard" then set the registry setting and see if you can force a memory dump on the working system.

most often with the memory dump someone can read it and get a pretty good idea what the problem is.


note: 2012 motherboard would be considered at its end of life.

 

[Da Rossa]

Right click it in device manager and see if you you can disable it.

Disabled the audio device which was failing to start. Still, can't know how that will help: if I no longer get BSODs, I still won't be able to tell if it was about this particular device.

if it does not create a memory dump, then something in the storage system is mess up.
(update the drivers from the motherboard vendors website)
then google "how to force a memory dump using a keyboard" then set the registry setting and see if you can force a memory dump on the working system.

Are you talking about this?
https://docs.microsoft.com/en-us/windows-hardware/drivers/debugger/forcing-a-system-crash-from-the-keyboard

In there, we see how to manually induce a crash. But how is that going to help? If I induce now, it will not capture the symptom I was looking to record, which happens only sporadically and can't be easily reproduced.

 

[johnbl]

problems with a computer still show up before the system crashes.
a mini dump will show the names of the drivers and the state of the CPU and bios versions. The windows debugger can check for corrupted windows files.

with a kernel memory dump the debugger can check internal error logs for errors and automated scripts can be run to check for common hardware and device driver problems.

it just saves time to do the memory dump and fix the common problems before you have to wait for a bugcheck to happen.

Right click it in device manager and see if you you can disable it.

Disabled the audio device which was failing to start. Still, can't know how that will help: if I no longer get BSODs, I still won't be able to tell if it was about this particular device.

if it does not create a memory dump, then something in the storage system is mess up.
(update the drivers from the motherboard vendors website)
then google "how to force a memory dump using a keyboard" then set the registry setting and see if you can force a memory dump on the working system.

Are you talking about this?
https://docs.microsoft.com/en-us/windows-hardware/drivers/debugger/forcing-a-system-crash-from-the-keyboard

In there, we see how to manually induce a crash. But how is that going to help? If I induce now, it will not capture the symptom I was looking to record, which happens only sporadically and can't be easily reproduced.

 

[Da Rossa]

problems with a computer still show up before the system crashes.
a mini dump will show the names of the drivers and the state of the CPU and bios versions. The windows debugger can check for corrupted windows files.

with a kernel memory dump the debugger can check internal error logs for errors and automated scripts can be run to check for common hardware and device driver problems.

it just saves time to do the memory dump and fix the common problems before you have to wait for a bugcheck to happen.

Sorry if I'm missing something, but I don't understand. Actually, I understand the part in which we can examine the minidump files to check what corrupt Windows files were to blame, or which hardware piece failed. But it's been four days since the last occurrence now. The minidump files for some reason are not in the C:\windows\minidump folder. How can I apply the above procedure to investigate that occurrence?

 

[johnbl]

- you can search your drive for *.dmp files
- if you run a cleanup tool like ccleaner it will delete the memory dumps.
- the user or OEM can disable the memory dumps or have them placed in a different directory.

there are certain bugs that prevent a memory dump from being saved. Bugs in anything related to storage (raid, BIOS, sata drivrs, bad cable connection to the drive, firmware bugs in solid state drives, not enough space on the drive, failure to have a pagefile) Make sure your system is saving memory dumps and not deleting them.

if no memory dump is being saved you have to figure out why. forcing a memory dump while the system is working helps to make sure everything is setup correctly.

problems with a computer still show up before the system crashes.
a mini dump will show the names of the drivers and the state of the CPU and bios versions. The windows debugger can check for corrupted windows files.

with a kernel memory dump the debugger can check internal error logs for errors and automated scripts can be run to check for common hardware and device driver problems.

it just saves time to do the memory dump and fix the common problems before you have to wait for a bugcheck to happen.

Sorry if I'm missing something, but I don't understand. Actually, I understand the part in which we can examine the minidump files to check what corrupt Windows files were to blame, or which hardware piece failed. But it's been four days since the last occurrence now. The minidump files for some reason are not in the C:\windows\minidump folder. How can I apply the above procedure to investigate that occurrence?

 

[Da Rossa]

- you can search your drive for *.dmp files

Didn't find anything but some Steam crash files. crash_csgo.exe_20171218232713_2.dmp

- if you run a cleanup tool like ccleaner it will delete the memory dumps.

Didn't run CCleaner since the occurrence.

- the user or OEM can disable the memory dumps or have them placed in a different directory.

Not my case, I've seen those files before. I also viewed them with BlueScreenView.

there are certain bugs that prevent a memory dump from being saved. Bugs in anything related to storage (raid, BIOS, sata drivrs, bad cable connection to the drive, firmware bugs in solid state drives, not enough space on the drive, failure to have a pagefile) Make sure your system is saving memory dumps and not deleting them.

if no memory dump is being saved you have to figure out why. forcing a memory dump while the system is working helps to make sure everything is setup correctly.

Now I get the point of it. I'll try.

...and I succeeded in inducing the manual crash! Just to have a look at the dump files. First, I had marked the option "complete memory dump" under Startup and Recovery under System Properties, but it was after I saw there were no .dmp files. I don't remember the option which was before, but it was much likely the default. I also changed the default save location to a handier place.

So I first created a full memory dump, just to realise it would simply create a file as big as my available RAM. Obviously I discarded it and changed to automatic, hit the combo keys again and, after restarting, this 887MB file was generated, and I believe BlueScreenView is not telling much:

Then I changed to "small memory dump (256kb)" and generated this:

... which seems to be more meaningful. Still, what can I learn from that?

 

[johnbl]

if you generate a dump, I would use the windows debugger to read it.
a minidump would only provide basic drivernames and bios info
a kernel memory dump provides what is running on each cpu, all of the driver data and internal error logs and kernel stack back traces with info about the hardware.

both can be useful but certain problems can only be debugged with the kernel memory dump.
(they are pretty big, 800 or 900mb would be pretty typical)

 

[Da Rossa]

I'm almost there: I changed the option in order to generate a kernel memory dump instead of an automatic or small dump. However, I can't find the native utility to read it. Articles found Googlewise suggest there is some dumpchk.exe, but there is no such program here. So how to 'officially' view the dump files on Windows 10?

 

[johnbl]

you would have to download the windows device driver development kit to get the tools. or now there is a new windbg.exe on the windows store (preview version, search for windbg in the windows store) that you can download and run.
better to put the memory dump on a cloud server like Microsoft one drive, share it for public access and post a link.
the kernel dump will contain information about the hardware and device drivers and some internal error logs. it would not contain user information. (other than the names of the running processes)

you can try the preview version of windbg but there are a lot of commands.
use this:
lmiftsm

this command will make a sorted list of the running drivers.

!sysinfo smbios
(will list the bios info and info about the cpu)

!for_each_module !chkimg @#ModuleName
(this will check each windows file for corruption, all 3rd party drivers will show a error)

!pnptriage
will show devices that are in the process of being installed (or can not be installed)

there are hundreds of various commands.

I'm almost there: I changed the option in order to generate a kernel memory dump instead of an automatic or small dump. However, I can't find the native utility to read it. Articles found Googlewise suggest there is some dumpchk.exe, but there is no such program here. So how to 'officially' view the dump files on Windows 10?

 

[Da Rossa]

you would have to download the windows device driver development kit to get the tools. or now there is a new windbg.exe on the windows store (preview version, search for windbg in the windows store) that you can download and run.
better to put the memory dump on a cloud server like Microsoft one drive, share it for public access and post a link.
the kernel dump will contain information about the hardware and device drivers and some internal error logs. it would not contain user information. (other than the names of the running processes)

you can try the preview version of windbg but there are a lot of commands.
use this:
lmiftsm

this command will make a sorted list of the running drivers.

!sysinfo smbios
(will list the bios info and info about the cpu)

!for_each_module !chkimg @#ModuleName
(this will check each windows file for corruption, all 3rd party drivers will show a error)

!pnptriage
will show devices that are in the process of being installed (or can not be installed)

there are hundreds of various commands.

I'm almost there: I changed the option in order to generate a kernel memory dump instead of an automatic or small dump. However, I can't find the native utility to read it. Articles found Googlewise suggest there is some dumpchk.exe, but there is no such program here. So how to 'officially' view the dump files on Windows 10?

Thanks, I downloaded WinDbg and was able to view the crash drump file. However, in order to create a dump file I had to induce the crash manually with Ctrl+2x Scroll Lock. Then I run all the commands you reommended me as examples. Would you like me to post them here so you can tell me what to look for when the real thing happens, or should I wait, because it won't be of much help?

 

[johnbl]

best to put them on a server like Microsoft onedrive, google docs or mediafire, share them for public access and post a link.

you would have to download the windows device driver development kit to get the tools. or now there is a new windbg.exe on the windows store (preview version, search for windbg in the windows store) that you can download and run.
better to put the memory dump on a cloud server like Microsoft one drive, share it for public access and post a link.
the kernel dump will contain information about the hardware and device drivers and some internal error logs. it would not contain user information. (other than the names of the running processes)

you can try the preview version of windbg but there are a lot of commands.
use this:
lmiftsm

this command will make a sorted list of the running drivers.

!sysinfo smbios
(will list the bios info and info about the cpu)

!for_each_module !chkimg @#ModuleName
(this will check each windows file for corruption, all 3rd party drivers will show a error)

!pnptriage
will show devices that are in the process of being installed (or can not be installed)

there are hundreds of various commands.

I'm almost there: I changed the option in order to generate a kernel memory dump instead of an automatic or small dump. However, I can't find the native utility to read it. Articles found Googlewise suggest there is some dumpchk.exe, but there is no such program here. So how to 'officially' view the dump files on Windows 10?

Thanks, I downloaded WinDbg and was able to view the crash drump file. However, in order to create a dump file I had to induce the crash manually with Ctrl+2x Scroll Lock. Then I run all the commands you reommended me as examples. Would you like me to post them here so you can tell me what to look for when the real thing happens, or should I wait, because it won't be of much help?

 

[Da Rossa]

Ok thanks John. I guess I'll have to wait for it to happen again. Thanks, I learned a lot with this topic!