PC Crashing

SPArton

SPArton

SPArton

Associate
Joined
13 Aug 2013
Posts
272
Hi guys,

I have been having major problems with my PC. It has been BSOD'ing with the "notskrnl.exe" supposedly causing the crashes. Running Win10 64bit. Happens at random times, usually when it's idle and I'm not using it. It will just lock and mouse/keyboard won't respond, leading me to need to force reset, or whilst I am using it the PC will crash to a BSOD.

I have replaced the RAM, PSU, GPU to no avail. Running an 840 Evo SSD, 4670k with a Z87-D3HP motherboard. Latest BIOS & SSD firmware.

Any ideas?
Thanks.
 
ntoskrnl BSODs can be somewhat common unfortunately. A lot of things can cause an ntoskrnl.exe BSOD. To list a few:

Corrupted Windows installation
Bad drivers (particularly video drivers)
Faulty RAM

Are you overclocking at all? If you are, reset everything to stock speeds and see if the BSODs persist.

What's your full system spec? If you've changed the RAM, PSU and GPU, it could just be a borked Windows installation - as in somethings corrupted Windows.
 
In power management do any of you guys let the usb or hard disks sleep? Or have you stopped all advanced power management options? My guess is some sort of power management issue.
 
Jameswa23 said:
I have just checked the settings, my HDD is currently set to turn off after 20 minutes of inactivity. I will not I have never changed this setting on any of my builds and was ot an issue up until 6 months ago. I would be surprised if it was this but i will give it a go.. I have nothing to loose at this point :)
Click to expand...

Go through all the power management settings and stop everything that wants to turn off after given times. Also stop is going into sleep and give it a see. I just have this feeling it's down to some sort of power management issue. Also stop it disabling usb devices in power management. I think its called selective suspend.
 
Jameswa23 said:
Ok, so an interesting turn of events. I put the 4 sticks back in and loaded into a game, after a couple of minutes my game itself crashed (this was one of the signs that i was having previously which could trigger a blue screen). So i took out the ram i tested the machine. just to try and make that clearer please see below

Board position A1 (first test with 2 sticks)
Board position B1 (Second test with other 2 sticks)
Board position A2 (first test with 2 sticks)
Board position B2 (Second test with other 2 sticks)

The machine seems to run fine when i have either of the two stick configurations running but start to play up once there are 4 sticks. This is now officially wrecking me head.

the ram I am using has always been in the motherboard from day one and was all bought at the same time, specs etc. I am beginning to wonder it it could be the board. possibly something to do with the interface between the A and B ram slots. this is heading into expensive territory.

Any suggestions?
Click to expand...

More voltage to the ram or the imc is required I recon. These things can just degrade over time and the imc is worked harder with more memory slots populated.
 
Jameswa23 said:
I think I have it. I have now got all 4 sticks of ram installed and seems to be stable. after a little trial and error i have found the a setup that appears to be stable.

I went into the Bios like you suggested and started plaing around with the timings and voltages on the ram setup. I knew from before that the 9-9-9-24 1333mhz setting was not working properly (this is the default for the ram when not in XMP mode), I also knew that the XMP setting (running at 9-9-9-24 1600mhz) was playing up when all 4 sticks were installed. I managed to narrow it down to 11-11-11-30 1866mhz which the system seemed to like but then had the undesired result of my system locking up after 40 minutes or stressing the system. i have finally settled for 11-11-11-35 1866mhz which seems to be running nicely. it will be interesting to see what this does overtime, hopefully eradicating the BSOD issue. these are all running at 1.65v

If I don't have a blue screen over the course of the next week that will be good.

Thanks for all of your help so far and i will keep you posted on how i get on (hopefully wont hear from me to soon).
Click to expand...

You will probably be able to get stable at the tighter timings with some adjustment to voltage of the IMC, I think back in the day of haswell that was the VCCA voltage:

Shamelessly stolen from tech power up:

VCCSA: Contains the memory controller, the PCIe controller, and other I/O domains. This is the main System Agent's domain voltage. The stock value sits somewhere between 0.800 V to 0.950 V, with most seen so far sitting between 0.850 V and 0.900 V.
Click to expand...

So a little bump in this voltage will likely see you get the stability at xmp.
 
Saw your post and gave it a try. had the same results with OCCT memtest and memtest 64. on another ryzen machine.

One of the 2 sticks is faulty. All Memory sockets work and the fault follows the bad stick around, For other on a fault finding mission, if the fault sticks to the same mem socket you could might need to try a new CPU/main board as the issue could be with the memory controller on the CPU or bad connections on the motherboard. .

Thanks Jameswa23
 
Not being funny but if you have not memtest86 how do you know it is not ram issue?

also you have had a number of BSOD, and you have persisted in using the system regardless without getting to the bottom of the cause. I suspect your windows is now corrupt. You may pass memtest86 but your OS has damaged files or drivers. I would strongly recommend a fresh install to test the stability of the system in order to isolate any hardware issues.
 
pc-guy said:
Not being funny but if you have not memtest86 how do you know it is not ram issue?

also you have had a number of BSOD, and you have persisted in using the system regardless without getting to the bottom of the cause. I suspect your windows is now corrupt. You may pass memtest86 but your OS has damaged files or drivers. I would strongly recommend a fresh install to test the stability of the system in order to isolate any hardware issues.
Click to expand...

That's the beauty of the running SFC /scannow from the command line. have occasionally had to re-install apps but haven't broken any windows system files yet.

Also dual or quad channel kits are great for diagnostics too. on my dual channel kit, I ran the memory on 1 stick . swapped them around and one clearly throws stuck address and other errors errors at the same address on the same test runs. the other come's back fine. you can swap dimm sockets and the errors only happen with the dodgy stick is in the machine. memtest64 ( and occt for that matter but with less detail ) throws the error almost right away or within the stuck address test Which is about within 2 or 3 minutes on my 32Gb kit, and in my case memtest86 pass half a day of testing 5-6 hours of testing.

I was fortunate to be able to borrow a similar slower kit from work. while waiting for an RMA. No errors. I didn't need the other kit to find the faulty ram module though.
 
Last edited:
I've had Memtest86 pass bad ram on multiple hour runs, but it did finally catch it as faulty on an overnight run. Replaced that ram and had a 100% stable system afterwards. I test all my new ram now on an overnight Memtest86 run.
 
You must log in or register to reply here.
Back
Top Bottom