[SOLVED] BSODs but NOT while gaming ?

[Subarashi]

Need some help figuring out what exactly is faulty so I can replace it. Personally I think it's either CPU or Mobo.

The problem
Issue is I get BSOD's but NOT while I am gaming and I also can get stutters both audio and video in certain situations and I can hear plops through speakers usually when I go from one app to the next. The most reliable way for me to get stutters and BSOD is to have a couple of browser tabs open and some apps and crucially to press on Windows Security in the hidden icons. Before this I had Norton antivirus but clicking on that could also give me the same issue so I uninstalled it thinking it was the culprit but then windows security gives me the same issues. Clicking on Windows security doesn't always lead to issues but definitely the most reliable way to create it. Other than that I get it mostly while watching videos online with multiple tabs open plus having window apps open and open task manager and other window apps.

What I tried already
At first I thought it was my GPU but I replaced my 2060 GTX super with my old 1060 of which I was almost certain it isn't faulty and I got the exact same issues still. Then I thought maybe it's my Ram so I did the windows memory diagnostic and memtest86 and got 0 errors. Also my SSD and HDD have no errors and are healthy. I also reinstalled windows twice full reset. LatencyMon Highest reported DPC/IRS is at like 500 usually and conclusion is system appears to be suitable for handling real-time audio and other tasks without dropouts. According to Hwinfo the heat on all components is fine at like 30-40 degrees idle maxing at like 60-65. All drivers are up to date. Driver verifier showed no issues other than lag when I am running all non microsoft drivers.

I have had multiple dump files and I viewed them with bluescreenview but I have little to no idea what it all means. But even though the bug checks are usually different the cause by address is always the same, ntoskrnl.exe+3f71d0. The last 3 bug check strings are KMODE_EXCEPTION_NO_HANDLED, IRQL_NOT_LESS_OR_EQUAL, and IRQL_GT_ZERO_AT_SYSTEM_SERVICE.

My latest dump file is below.
Kernel dump file --> https://drive.google.com/file/d/1JoT0FJU2ybmXgE9JtkBYkQNi3-V-vMRA/view?usp=sharing

System (8months old)specs are:
Windows 10 Home
BaseBoard Manufacturer MEDION
BaseBoard Product B550A4-EM
Processor: AMD Ryzen 7 PRO 4750G with Raedon Graphics
GPU: RTX Nvidia 2060 super (testing with 1060 atm)
RAM: 32GB Samsung M378A2G43AB3-CWE
SSD: E12-512G-PHISON-SSD-B27-BB1
HDD: TOSHIBA DT01ACA100
PSU 650W not sure about manufacturer
Some older logitech speakers and philips Monitor.
Let me now if I you need more info!

 

[Colif]

PSU 650W not sure about manufacturer

can you find out what make/model it is

Some PSU have issues dealing with Ryzen at idle. What Ryzen does is it has a setting called Rush to idle that can trip up PSU. When Windows has a task that needs a core Ryzen will wake the core, boost up to the speed needed to do operation and then when its finished, it will rush back to 0 again and its this constant rush up and down that can trip over PSU that aren't designed for it.

Aldi PC... I recognize Main board brand. Wonder what PSU brand they use.

I will ask a friend to convert the dumps but if any part of the PC is likely

Before I blamed CPU & MB I would look at PSU.

 

[Subarashi]

can you find out what make/model it is

Some PSU have issues dealing with Ryzen at idle. What Ryzen does is it has a setting called Rush to idle that can trip up PSU. When Windows has a task that needs a core Ryzen will wake the core, boost up to the speed needed to do operation and then when its finished, it will rush back to 0 again and its this constant rush up and down that can trip over PSU that aren't designed for it.

Aldi PC... I recognize Main board brand. Wonder what PSU brand they use.

I will ask a friend to convert the dumps but if any part of the PC is likely

Before I blamed CPU & MB I would look at PSU.

Hi the Power supply is "InWin A65 650W" if I can believe their website. Couldn't find this info on the PSU itself even when I open the case up.
Also I have to add the pc had been having stutter issues from the get go when I bought it. But they were not frequent and very short. I would get like 2-3 audio and video stutter seemingly out of the blue. Now in hindsight I should have put more importance on it and find out exactly what caused them. The only reliable way for me to do it now is to open Window Security after I for example am watching a video on youtube. It could very well be what you described with the rush to idle. Is there a way for me to turn that rush to idle feature off or some other way to mitigate for it with software/bios other than buying a new PSU or other hardware?

 

[Colif]

https://www.in-win.com/en/gaming-power-supply/list/427
80 plus white

it doesn't surprise me that Medion would use it. Its not the best PSU but it might not be cause.

I will see what dumps show before jumping to conclusions.

 

[TheOldieRNZ]

I had a pretty similar problem awhile back with my old AMD system. First it was BSODs. Cleaning up your system (and fresh new windows install) reduced the occasion, but it still happens. More often when under load. Then it no longer BSODs and kernel failures or anything, but straight up to total freeze and lockups. Like your screen just frozen in time and nothing you can do but to cycle the power (hard reset). Event Viewer only gives "Previous system shutdown at blah3x is unexpected" means it didn't even received any error and regarded it as a power failure (i.e: blackouts or power lost).

Tried everything, change/swapped every parts (but the mobo and cpu and psu). Use different HDDs, use SSDs, clean install on a freshly re-initialized disk, clean install on brand new SSD. Freezes still happens (but almost never a BSOD anymore, only occasionally happened when I messed up some custom made hardware drivers). I have a hard time believing that it is my PSU, after all it is a highly sophisticated brand (Seasonic) and also one of their best (if not THE best in the world) type: the Seasonic Platinum series (also expensive as hell here, you can literally build a full low spec rig from the cost of a single PSU). And since I'm so blinded by it, I never bothered to try replacing it, till I had to repair a friend's PC and on a whim, tested his PSU (just the usual low quality, low grade 450W ATX PSU that came with the trashy PC case, here in my country you can get new one for as low as $5 (which is why they usually being sold in bulk).

It never freezes when I'm using that low quality PSU. No freeze, no BSOD, everything ran smoothly. Curious, I trashed everything on it: Prime95, AIDA64, OCCT, Unigine (sorry no Cinebench, its an old rig from early 2000s). It passed with flying colors. No errors, no BSOD, no freezes.

Try replacing your PSU (or just borrow one for testing), who knows you're having the same issue as I am.

 

[Colif]

It could very well be what you described with the rush to idle. Is there a way for me to turn that rush to idle feature off or some other way to mitigate for it with software/bios other than buying a new PSU or other hardware?

It is a feature of how the CPU work. You can't stop it. All Ryzen CPU do it by design.

Wait and see what dumps show

I wouldn't put a $5 PSU anywhere apart from a compactor to make sure its destroyed and no one thought it might be useful. I have gone from buying Cases that came with no name PSU to choosing the PSU as 1st item in my next PC. Priorities change when you realise its heart of system. If its bad, it can make everything else bad as well. It can kill parts. Bad PSU killed 3 HDD I owned in a 6 month period and for a long time I blamed the HDD, now I know better. I have had less parts die on me since I started using good PSU.

If Its a Seasonic Plat, didn't it have a 10 year warranty? Why not rma it if it was cause of problems. I would do that before using a cheaper PSU...

 

[TheOldieRNZ]

It is a feature of how the CPU work. You can't stop it. All Ryzen CPU do it by design.

Wait and see what dumps show

I wouldn't put a $5 PSU anywhere apart from a compactor to make sure its destroyed and no one thought it might be useful. I have gone from buying Cases that came with no name PSU to choosing the PSU as 1st item in my next PC. Priorities change when you realise its heart of system. If its bad, it can make everything else bad as well. It can kill parts. Bad PSU killed 3 HDD I owned in a 6 month period and for a long time I blamed the HDD, now I know better. I have had less parts die on me since I started using good PSU.

If Its a Seasonic Plat, didn't it have a 10 year warranty? Why not rma it if it was cause of problems. I would do that before using a cheaper PSU...

Oh that was just for testing purposes, and unfortunately the company who used to distribute/sell it here long closed down and gone. And the new distributor refused to accept it (I did contacted them). Also its not 10 years on that model (something2 ss 760 xp2 something, already dismantled, scavenged and gave the leftovers to a friend of mine who ran a hardware repair shop), only 1 year from the store (closed) and additional 2 years from the old authorized distributor (also closed).

It was bought in 2013 btw. Seasonic International said that I need to sent it to them first, and since its quite heavy, shipping (International Shipping to Hong Kong, their Asia HQ) cost for it is quite heavy too and even then they can't promise an RMA (since the model is also already long EOL). Best they can do is repair (which is NOT free) and I have to handle all shipping costs (for the return too). The shipping cost to HK alone already surpassed the price I paid buying it back in 2013, so I choose not too.

Also yes, the first thing I did when buying new (non branded) pc case, is to pluck the PSU and throw it into the trash bin. I've been a hardware technician for several ISPs and Datacenters alongside (used to be) an enthusiast (I'm 50 years old and retired (8088/8086 gen)). PSUs are at the very top of my list (thus why I bought that Seasonic Platinum all those years ago when the majority of enthusiasts (cough overclockers) never even heard or knew what 80+ standard is, when chaining ATX PSUs and DIY Molex adapters still considered 'the norm').

I'm currently using a custom build, modified 750W PSU from Aerocool. It's an 80+ Gold, and quite cheap (about $70 (old stock discount), same model but now with fancy RGB is around $110 locally), dismantle it on arrival and replaced everything I can with higher grade parts (nichihon solid caps, higher grade mosfets, better voltage regulators, higher thermal and amp diodes/bridges from TSMC, even replaced some aluminum heatsinks with copper ones and also replaced the crappy fan with a SANYO DENKI 120mm one). I'm not able to actually tested it due to the outbreaks, unable to go to my workshop (where the oscillators and the precision measuring equipment are), but my DMM shows that everything's good, even under full 750W load the voltage only dropped for like 0.05V on the 12V (its a single rail design) and that is pretty much all I can test it with what I have at home.

 

[Subarashi]

@TheOldieRNZ I don't appreciate the high-jacking of this post. I am trying to get my pc issue diagnosed, if you want to have a discussion with him please private message him or open up your own post.

 

[Subarashi]

@Colif Another BSOD here is the kernel Dump ---> https://drive.google.com/file/d/1-2xeEc1J39FhsQfx2SvRrDCryUKwww1e/view?usp=sharing

INTERRUPT_EXCEPTION_NOT_HANDLED, this time with bug check code 0x0000003d

 

[Colif]

my bad, I did ask him about PSU

my friend is in a different time zone to me so he will reply when he can

 

[Subarashi]

@Colif I want to add that I did a Realbench 2.56 stress test which put my cpu usage at 100% for 15 minutes went up to about 80 degrees heat plus my gpu was running at about 80% and I did not get any error or bsod. So at this point I am not suspecting my cpu, but am thinking it could be either my psu, mobo, some driver interaction or my ssd or hdd. I think the fact that when I am not doing much and suddenly press on an antivirus program (even just windows defender) that causes stutter or bsod is some kind of hint. I am not sure what though.

 

[Colif]

my friend has been busy the last few days so I hope he is in soon.

then I can look at crashes

IRQL_GT_ZERO_AT_SYSTEM_SERVICE.

I don't see this one a lot. rarely in fact.

but am thinking it could be either my psu, mobo

the 2 hardest parts to test without spares. PSU you could at least use a multi meter to test, there are no tests for MB.

PSU
the paper clip method - https://forums.tomshardware.com/threads/what-is-the-paperclip-method-of-testing-a-psu.1336402/
or multimeter,
or in the BIOS to check the +3.3V, +5V, and +12V. - https://www.lifewire.com/power-supply-voltage-tolerances-2624583

Most of the time people have problems with PSU when its under load. Yours appears to be idle so looking at it in bios might show something

you have HWINFO installed.
In the sensors screen, along bottom are icons. Click on the icon to right of the Clock icon, it says "logging start" if you hover mouse cursor over it.
This opens file explorer, create a log file and run it while you idle or until you crash.
You will need to start it each time.
The results are saved as CSV files that can be read in excel or Google docs.

If you want help looking at results, upload to a file sharing website and show link here.

 

[Subarashi]

my friend has been busy the last few days so I hope he is in soon.

then I can look at crashes


I don't see this one a lot. rarely in fact.



the 2 hardest parts to test without spares. PSU you could at least use a multi meter to test, there are no tests for MB.

PSU
the paper clip method - https://forums.tomshardware.com/threads/what-is-the-paperclip-method-of-testing-a-psu.1336402/
or multimeter,
or in the BIOS to check the +3.3V, +5V, and +12V. - https://www.lifewire.com/power-supply-voltage-tolerances-2624583

Most of the time people have problems with PSU when its under load. Yours appears to be idle so looking at it in bios might show something

you have HWINFO installed.
In the sensors screen, along bottom are icons. Click on the icon to right of the Clock icon, it says "logging start" if you hover mouse cursor over it.
This opens file explorer, create a log file and run it while you idle or until you crash.
You will need to start it each time.
The results are saved as CSV files that can be read in excel or Google docs.

If you want help looking at results, upload to a file sharing website and show link here.

I couldn't find the voltages in the bios but I will look again, the reading on Hwinfo seems off though....

 

[DSzymborski]

Yeah, software display of measurements is frequently quite off. Your PC wouldn't be running with those voltages on the rails.

I think I would swap out the PSU just to definitely eliminate it as a source of problems, given that it's not really an acceptable level of quality anyway.

 

[Colif]

HWINFO isn't always accurate.

hope those aren't right
since all 3 are way below what they should be.

I would get a PSU from someone like EVGA or Corsair and it should handle the idle times better

 

[gardenman]

Both links that you uploaded are shortcuts to a memory.dmp file. A shortcut won't work. I'd prefer minidumps rather than the full memory.dmp. Copy the files from C:\Windows\minidump to your documents folder first. Then upload them, and provide a link.

 

[Subarashi]

Both links that you uploaded are shortcuts to a memory.dmp file. A shortcut won't work. I'd prefer minidumps rather than the full memory.dmp. Copy the files from C:\Windows\minidump to your documents folder first. Then upload them, and provide a link.

Yeah that was really stupid on my part hopefully this works it's a kernel minidump https://drive.google.com/file/d/1K6s6u_sHAIWbmNoB9sDyJOeWSyOGUMDT/view?usp=sharing
Since it overwrites I only have this one now, I turned that setting off this time.

 

[gardenman]

The full memory.dmp that you posted will take me at least 2 hours to download. Most of the time memory.dmp files are corrupted. Either during the save/crashing process, or the uploading or downloading process. I do want to help you if possible, but I'd prefer minidumps.

 

[Subarashi]

The full memory.dmp that you posted will take me at least 2 hours to download. Most of the time memory.dmp files are corrupted. Either during the save/crashing process, or the uploading or downloading process. I do want to help you if possible, but I'd prefer minidumps.

Unfortunately I don't have a minidump as you say. Is it the small memory dump you want? If so i'll choose that option and wait for the next bsod. Also for some reason automatic memory dump never produces one which is why I did a kernel memory dump.

 

[gardenman]

Yes, a small memory dump (this creates a file in the C:\Windows\minidump folder).

 

[Subarashi]

@gardenman Hopefully you could still look at these for me.

https://drive.google.com/file/d/1cCnlK4g-msWH4jRAIPtFx4cGyc-dpgvj/view?usp=sharing

https://drive.google.com/file/d/1WoL3xxbCOG1mLXclho3gZ1n32JPgj1gM/view?usp=sharing

https://drive.google.com/file/d/1CbS_JwdlfYM7LwiKgsx1bTTzOhbgBOp5/view?usp=sharing

https://drive.google.com/file/d/1w2-lAVt0UrhITxjU1L1RnwCy0_Nx-ck7/view?usp=sharing

 

[gardenman]

I ran the dump files through the debugger and got the following information: https://jsfiddle.net/jaevft6r/show This link is for anyone wanting to help. You do not have to view it. It is safe to "run the fiddle" as the page asks.

Spoiler: Dump File Results

File information:	092521-9796-01.dmp (Sep 25 2021 - 07:25:14)
Bugcheck:	IRQL_NOT_LESS_OR_EQUAL (A)
Probably caused by:	hardware (Process: System)
Uptime:	0 Day(s), 1 Hour(s), 55 Min(s), and 00 Sec(s)

File information:	091221-7359-01.dmp (Sep 12 2021 - 13:01:54)
Bugcheck:	INTERRUPT_EXCEPTION_NOT_HANDLED (3D)
Probably caused by:	memory_corruption (Process: System)
Uptime:	0 Day(s), 7 Hour(s), 14 Min(s), and 58 Sec(s)

File information:	091121-8640-01.dmp (Sep 11 2021 - 14:19:16)
Bugcheck:	KMODE_EXCEPTION_NOT_HANDLED (1E)
Probably caused by:	memory_corruption (Process: System)
Uptime:	0 Day(s), 8 Hour(s), 13 Min(s), and 23 Sec(s)

File information:	091021-6500-01.dmp (Sep 10 2021 - 07:59:14)
Bugcheck:	IRQL_NOT_LESS_OR_EQUAL (A)
Probably caused by:	memory_corruption (Process: System)
Uptime:	0 Day(s), 0 Hour(s), 14 Min(s), and 28 Sec(s)

Comment: hardware was mentioned in the first crash.

Colif (or anyone else) will reply with more info later based upon this info. This information can be used by others to help you. Please wait for additional answers.

Edit: Results for dump below: https://jsfiddle.net/e9u1msg7/show

Spoiler: Dump File Results

File information:	092621-9250-01.dmp (Sep 26 2021 - 09:24:06)
Bugcheck:	DRIVER_IRQL_NOT_LESS_OR_EQUAL (D1)
Probably caused by:	ntkrnlmp.exe (Process: System)
Uptime:	0 Day(s), 3 Hour(s), 06 Min(s), and 09 Sec(s)

 

[Subarashi]

Here is another minidump from a bsod of today.
https://drive.google.com/file/d/1KQxKCnxpKkXfPAu_YLc3lUhMpTmF6tw_/view?usp=sharing