Windows 10 BSOD wierdness

Q

Quartz-1

Supreme [H]ardness
Joined
May 20, 2011
Messages
4,257
Windows has been BSODing on me recently. I've had Memory Management twice, and Invalid Push Lock Flag and PFN List Corrupt BSODs once each, perhaps one crash every other day. This should point to RAM but I've tested the RAM and it passes just fine. CPU temperatures are fine. There appears to be no rhyme or reason for the crashes. The PC is on all day every day. The crashes started before I got my Astro A50 headset.
 
What happens when you run a Prime 95 stress test? Is anything overclocked? Have you already reset your bios so that everything is at stock?
 
Well I haven't messed with windows a ton lately but seeing as the windows folk aren't offering much.

You seem to be pointing to this issue being related to your recent headset addition.

So I would start there....

Ensure its plugged into USB 3

Check and make sure you have the newest chipset drivers for your MB usb chipsets.

Uninstall and reinstall your usb headset. and reboot.

Once your rebooted check your device manager and ensure everything seems fine there.

If your still getting BSOD.... then rule out the headset by simply uninstalling it and seeing if that solves the issue. If not you can look elsewhere. If not then perhaps you can check out the manufacturers sources... perhaps they have a forum ect. First try and narrow it down to that bit of kit.

If you do narrow it down to the headset.... perhaps see if its some sort of dual audio driver issue by perhaps disabelling your main audio card. (it sounds like your headset uses usb audio if I'm wrong ignore me) if its on board audio you should be able to disable it in your bios.
 
ChadD said:
Well I haven't messed with windows a ton lately but seeing as the windows folk aren't offering much.

You seem to be pointing to this issue being related to your recent headset addition.

So I would start there....

Ensure its plugged into USB 3

Check and make sure you have the newest chipset drivers for your MB usb chipsets.

Uninstall and reinstall your usb headset. and reboot.

Once your rebooted check your device manager and ensure everything seems fine there.

If your still getting BSOD.... then rule out the headset by simply uninstalling it and seeing if that solves the issue. If not you can look elsewhere. If not then perhaps you can check out the manufacturers sources... perhaps they have a forum ect. First try and narrow it down to that bit of kit.

If you do narrow it down to the headset.... perhaps see if its some sort of dual audio driver issue by perhaps disabelling your main audio card. (it sounds like your headset uses usb audio if I'm wrong ignore me) if its on board audio you should be able to disable it in your bios.
Click to expand...
Quartz-1 said:
The crashes started before I got my Astro A50 headset.
Click to expand...
why the snide "windows folk" bs? he got one reply and then your misguided input...

OP. as MOG asked, any OC? if so, reset everything to stock to eliminate that first. if it stops then your oc has degraded and you need to fiddle with that. maybe a bit lower or a bit more voltage.
you could also try bumping the RAM voltage a bit.
is there anything else that was changed or added before the crashing started? driver updates, background app etc.
also test your cmos battery. if its getting low it can cause all sorts of weird problem.
and disconnect/remove any thing that you can go without until you figure it out.
bios and drivers up to date?
good luck!
 
pendragon1 said:
why the snide "windows folk" bs? he got one reply and then your misguided input...
Click to expand...

Fair enough I did read that before as after for some reason. (I guess my mind just thought why mention his headset if he didn't suspect it lol) I guess that is why its always a good idea to read things twice.

Having said that... reading again lists one BSOD reporting a PFN list error. (page frame list) as well as normal mem mang errors. So ya your advice of checking ram settings is solid. Also the bios bat, sure dying bios bats can cause all kinds of strange stuffs. Considering the page frame error though if everything seems to check out with the battery and your device drivers all seem to be up to date and running... consider your software as well. Make sure your running one virus scanner (and that its up to date and your system is clean.. as page file issues can be caused by messed up registry entries and malware) also look at anything you installed recently that may have perhaps made some odd reg changes. Although ya PFN errors mostly point to ram so for sure check that as pendragon suggests, also there is also a smaller possibility its related to a hard drive that is failing somehow. I would run some disk health tests on the drive hosting your swap. I notice your sig says your running a 1 TB SSD I would find a util to determine its health most of the major brands offer some form of s.m.a.r.t. monitoring util.
 
Why not just go through event viewer?

Perhaps take a look at Reliability Monitor while you're at it?
 
I would test the machine with another PSU as the first thing. Flaky power can and will cause the weirdest symptoms and crashes. Previously known good overclock not working anymore is typical for having power fluctuations.

I had a highly OC:d machine that little by little became more and more unstable. It also hesitated sometimes when trying to boot it. I had to drop the overclocks gradually to keep it stable. When it reached the point that even stock clocks were not working anymore I realised I have to start troubleshooting. Despite my suspicions of having fried the CPU or ram, I monitored voltages on the computer and saw that they fluctuated. Not by much but still. So I ordered a new PSU and lo and behold the computer was again stable to a full OC.
 
Thanks for all the replies.
Prime95 worked fine testing the RAM. As noted in my OP.
There's no overclocking - my CPU is a 3770S which doesn't overclock.
One of the crashes did cause some disk corruption, and resulted in a sea of red in the event logs. At the time I ran CHKDSK /F which appears to have fixed that; time will tell.
Crystal Disk Info reports that both SSDs are in good condition.
The PSU is under 3 years old, is a quality model, and is behind an inline UPS.

Reliability Monitor shows problems with Excel, Edge, and Treesize Pro. The last simply crashes upon startup.

I'm thinking that it might be time for a therapeutic reinstall. I need to rebuild the server too.
 
Quartz-1 said:
Thanks for all the replies.
Prime95 worked fine testing the RAM. As noted in my OP.
There's no overclocking - my CPU is a 3770S which doesn't overclock.
One of the crashes did cause some disk corruption, and resulted in a sea of red in the event logs. At the time I ran CHKDSK /F which appears to have fixed that; time will tell.
Crystal Disk Info reports that both SSDs are in good condition.
The PSU is under 3 years old, is a quality model, and is behind an inline UPS.

Reliability Monitor shows problems with Excel, Edge, and Treesize Pro. The last simply crashes upon startup.

I'm thinking that it might be time for a therapeutic reinstall. I need to rebuild the server too.
Click to expand...

I had a Seasonic platinum PSU that went bad 3 days from purchase. Check the PSU if you can.
 
Quartz-1 said:
Unfortunately I don't have a spare PSU of sufficient wattage.
Click to expand...

Most people have way oversized PSU:s. If you have, say, a 450W PSU and not some monster hungry graphics card you could try it just to see if your BSOD is gone.
 
rezerekted said:
Always good to have a backup around so if you can afford it get another PSU.
Click to expand...

B00nie said:
Most people have way oversized PSU:s. If you have, say, a 450W PSU and not some monster hungry graphics card you could try it just to see if your BSOD is gone.
Click to expand...

Look for a sale on a reasonable PS. As BOOnie points out a min to keep things powered makes a good backup. Finding a sale on a mid range powered PS isn't often hard. It doesn't have to be the top of the line this is my PS for 10 years type unit. Just keep it around so if things shit the bed you can swap it in for a few weeks or so till you can pickup a proper replacement. I have kept a few clients junker PS around over the years and they have come in handy a couple times, nothing would suck more then a dead PS and waiting for an online purchase to show up. :) I would suggest checking your bios ram settings.... sometimes the auto settings seems to get things wrong. Perhaps set your ram manually and see if that helps, you could even kick the voltage up a bit and test that.

Having said that though if you have software that is crashing on start... you do seem to have some sort of software issue. It may have been caused by the crash though of course. Meaning you sort of have a couple issues instead. I know windows repair tools are mostly BS still perhaps its worth a try before you go the full reinstall route. If you do go the full reinstall route and the issue seems solved perhaps make a fresh back up with clonezilla or something like it.
 
Install the Windows Debugging Tools from here: https://go.microsoft.com/fwlink/p/?LinkID=845298
Which installs to C:\Program Files (x86)\Windows Kits\10\Debuggers\x64\windbg.exe

File > Open Crash Dump > C:\Windows\minidump\
In the window, click on or type in: !analyze -v

Look for the output line that say "probably caused by: x"
You have to reopen the program to check each minidump.

Make sure your drivers and firmware are up to date.
 
I always keep a spare, brand new 450 watt quality PSU on hand for diagnostic purposes. At the end of the day I usually end up selling it anyway.
 
I've got spare PSUs. I just don't have one capable of powering a pair of Titan Xs.
 
I've been in similar situations before. My issue ended up being a bad stick of RAM. I could not get a Memtest of any kind to find it but I kept on having memory related issues. One day I just screw it and bought two new sticks of RAM. I replaced the expensive Corsair set that was in there with some G.Skill set that was on sale. My son is running that system now, same G.Skill ram, zero issues. I still have the Corsair set in a box in my desk. I have no idea which stick is bad, I cannot get it show up on any memory tester.
 
Just as an FYI memtest is kinda of like smart on HDs, just because it doesn't error doesn't mean you don't have a problem, its just if you do error there you probably do have a problem . Only way to be sure is start yanking sticks. I've had everything go bad on a computer at some point in time, but for me the most likely problem is the ram (i'm including slots going bad on the mb), followed by usb, followed by PSU. I've had plenty of graphics cards go bad over the years, but i'd test ram first, graphics next since its the easiest and i wouldn't rely on software to test it. empirical testing is significantly more reliable.
 
-PK- said:
Look for the output line that say "probably caused by: x"
You have to reopen the program to check each minidump.
Click to expand...

A look at memory.dmp reports 'Probably caused by : memory_corruption ( nt!MiDeleteValidSystemPage+120da2 )'

A look at the minidumps reports 'Probably caused by : Unknown_Image ( ANALYSIS_INCONCLUSIVE )' Fat lot of use that is! :)
 
Reliability Monitor, quite often it gives clues that aren't anywhere near as cryptic.

Click Start > begin typing 'reliability' without the quotes and click the first result in the list.
 
If it starts up again, try pulling one of the Titans out.
If it runs ok, swap the other one in.

If they both run well separately but not together, then I'd guess PSU issues.

It would bother me to have an unstable machine just magically become stable again.
You can be fairly certain there's still a problem.

.
 
You must log in or register to reply here.
Back
Top