Boot Failure Detected. PC occasionally reboots to Bios, Event 41 Kernal shows up in Event viewer? New W10 install, new PSU, new SSD's.

[sastusbulbas]

Ok so I am starting to get a little frustrated with my lads PC.

He has had his PC for a good few years, a few months ago we changed his memory from Corsair Vengeance DDR3 2400 16gb to Corsair Dominator DDR3 2132 and ran his Windows 7 PC at that for a few months no issues, memory was only changed due to the Gold Vengeance clashing whereas the Dominator has Black/Orange light bars that went with his Black/Orange motherboard.

But we were supposed to be migrating him to Windows 10, and downsizing his PSU, so we did that two weekends ago, but since upgrading around 4 times a week we get random reboots to BIOS with the Boot Failure Detected screen.
Every time it reboots I end up either running it all at defaults or having to manually reset memory, thats it. It will still occasionally reboot back to boot failure detected screen at bios regardless of what I change in BIOS.
On top of that, on his 1st evening his wireless went loopy and stopped working, he said his stick had got rather too warm, then W10 refused to connect with it, TP-Link 3200 and it looked like a few had issues with that USB adapter and I thought no more about it and was planning to get a new PCI Wireless adapter.

When he was on Windows 7 nothing ever went wrong, it just ran problem free even with the TP-Link on 5G networking. We then decided a dust and W10 upgrade was due, and that my old 1000w PSU could come out, we added a brand new NZXT C750 psu, that ran on a bench with two other W10 builds for a couple of weeks with zero issues prior to going into his PC which it was intended for, we also added a new pair of SSD's, PNY 250gb and 960gb for OS and games as we were changing to W10 and wanted a fresh install. My two other kids have the same drives and W10 with no issues.

At 1st I thought some settings to do with sleep and power, so changed them, no change as still an occasional boot failure, then device compatability but nothing in device manager flagged, even the TP link was showing as no problems despite W10 refusing to work with it, then thought CMOS or CMOS battery, so the CMOS was cleared and a new battery fitted, then thought wiring, all wiring is fine, GPU drivers up to date, ran Memtest from a USB and even with XMP the memory passed, checked drives both 100% fine on CrystalDiskInfo, nothing shows as having a problem, nothing is overclocked. Temps idle around 30 and not seen them pass high 40's. Feel like I am overlooking something simple here, and event viewer shows the same thing every time it boots to that Boot Failure screen. Event 41.

Event Viewer shows 4 criticals this week that tie in with the reboots, it is usually when the system is switched on, happened once during a game, though it may have been an update doing that. Event 41 Kernel Power is what the event states. And my lad is asking is it because we changed the PSU. Any opinions or experience of this would be welcome. The PSU is a brand new NZXT C750 which I dont think are bad units.

I have read others have had this issue and it was OS related due to some bug. Nothing to do with the Power Supply as some found out after replacing theirs? Tempted to again clear CMOS, do a fresh Windows USB install with an up to date ISO and only th esingle SSD installed. The PC runs nicely apart from that occasional reboot, but I feel my lad is now unhappy switching it on in case he gets that reboot and I can see he is now Disappointed with his PC after years of wanting nothing changed in it apart from wanting W10 and asking for 32gb of black memory as he hated the Gold colour and why does every kid think they need 32gb of ram.



His system. (Windows Boot Manager has boot priority before both drives in bios if that makes sense?) I did have the full PC built with both SSD in place when I installed an old version of Windows 10 from a USB drive too. Only updated GPU drivers and Windows and added iCue.

Windows 10 Pro 20H2 (Not activated as we want the reboot resolved 1st)
Intel 4770K at stock
CoolerMaster Nepton 240M CPU cooler as intake
Gigabyte Z97X SOC
Corsair Dominator DDR3 16gb 2133C9
PNY C900 250gb SSD
PNY C00 960gb SSD
Powercolor RX570 8gb GPU
NZXT C750 Gold Modular PSU
NZXT S340 case with two NZXT case fans as exhaust
Corsair K63 Keyboard
Corsair M65 Pro Mouse
D-Link USB Network adapter

I want to get this system back to 100% stable as it was previous, as all my kids will most likely be on 4770k with 16gb ddr3 and W10 for the next couple of years.

 

[mickyflinn]

Random crashes are a pain .

Do you have you old psu to try ?

 

[Tetras]

I find your description kind of hard to follow, because it sounds like you've made a lot of changes to the PC and I can't really establish a timeline, but if I boil it down to basics: everything run fine until you:

- changed the PSU

The memory change seems a red herring because the system was stable for several months?

Since then, you have:

- installed Windows 10
- added two new SSDs
- having problems with the wireless stick

When the PC reboots it goes into a screen saying:
- boot failure detected (this is the motherboard, right?), which suggests it tried to boot and failed?

So, my logic goes like this:
- the memory change worked and is stable, no problem
- the PSU did it (or something related, like a dodgy power cable or something is loose inside the PC, due to the PSU change)
- something went faulty at the same time and the PSU is a red herring

"boot failure detected" - I have three theories to what this message relates to
- SSD has disappeared so the PC can't boot
- the system is not actually stable, therefore it struggles to boot (e.g. now having 4 sticks can cause this, if the voltage is too low, but you said it was stable for months)
- the motherboard detects that power has been lost and so assumes the system was not stable due to OC, or something along those lines

 

[Chris Beard]

I think that everything is fine hardware wise but the new PSU has slightly greater v-droop or just a lower 12v line than the old one.

I'd suggest stability testing with Prime95. If you're not stable try adjusting the voltages for CPU, Socket and Memory within safe limits one-by-one to tune in which one(s) need to be increased with the new PSU.

 

[Quartz]

Is the PC on a UPS? It may be that the new PSU is more sensitive than the old one.

 

[sastusbulbas]

I find your description kind of hard to follow, because it sounds like you've made a lot of changes to the PC and I can't really establish a timeline, but if I boil it down to basics: everything run fine until you:

- changed the PSU

The memory change seems a red herring because the system was stable for several months?

Since then, you have:

- installed Windows 10
- added two new SSDs
- having problems with the wireless stick

When the PC reboots it goes into a screen saying:
- boot failure detected (this is the motherboard, right?), which suggests it tried to boot and failed?

So, my logic goes like this:
- the memory change worked and is stable, no problem
- the PSU did it (or something related, like a dodgy power cable or something is loose inside the PC, due to the PSU change)
- something went faulty at the same time and the PSU is a red herring

"boot failure detected" - I have three theories to what this message relates to
- SSD has disappeared so the PC can't boot
- the system is not actually stable, therefore it struggles to boot (e.g. now having 4 sticks can cause this, if the voltage is too low, but you said it was stable for months)
- the motherboard detects that power has been lost and so assumes the system was not stable due to OC, or something along those lines

Sorry I thought the timeline was clear, he has used the PC for years with no change, a few months ago we changed his memory purely on cosmetics, zero Event 41 issues in his Windows 7, we simply set memory timings and it never batted an eye.

A few months later IE a couple of weeks back, the next change is a bit more involving as we were going from Windows 7 to a fresh install of Windows 10, while adding 2 new SSD and a new PSU at the same time.

Regarding the TP Link TL-WDN3200 USB adapter, a quick google shows many have had issues with that USB adapter and Windows 10. A D-Link adapter is in the same port no issues.

Boot Failure Detected initially had me wonder hardware or PSU, hence me going through Bios setings, CMOS, Memtest, swapping ram slots, monitoring temps. GPU driver. Only looked at Event Viewer yesterday.


So PC was stable and never rebooted to bios on W7 with the old PSU, old SSD's and Windows 7, so this Kernel issue comes with the recent change of PSU, two SSD, and Windows 10 from my deductions so far.

The new PSU was used to test build two other PC's before going into my sons build. Those test builds went fine, both were tested with the same W10 USB install my lad got. PSU is brand new retail bought for my lads PC as he had an overkill 1000w model from my dual GPU days.

As the reboot is so random and every time apart from once been at power up, and I am usually working, I only check and try to change one thing a a time. Was two days before it last happened yesterday.

Interestingly if you Google error 41 kernel power, it seems a common problem and there to be no straight answer with many seemingly changing hardware for nothing and it turning out to be a Windows 10 bug. I have never seen the issue. Hence asking here. Loads of advice online much of it settings due to Windows 10.

I mentioned I had changed power settings, I revisited that a little deeper in advanced settings, made sure HDD was never off and changed its Power Off button to actually off as described in one article related to this issue as I was still on fast boot. So will see if this helps for the next few days.

Some options I am thinking of as next steps, re-install all drivers, put his old SSD's with W7 back in and see if it re-occurs, try a fresh clean W10 install with just the basics installed on a cleared CMOS, or change the PSU. Was hoping it was an issue someone else may have had experience with. Right now after looking at the internet stories I feel its a W10 issue.

 

[Tetras]

I could be wrong, but from the sound of what you're saying, I think the critical power events just reflect the power loss?

If they happen prior to the power loss then I'd consider that more significant.

 

[therealdippy]

Event 41 is simply that Windows did not shut down properly - you'll get that if you simply power off a PC - after Windows restarts that event will be in the log.
I assume that "Boot Failure Detected screen" means that POST fails? If so this is a HW issue, either an actual fault or a configuration issue.
I wonder about your comments "having to manually reset memory" and "simply set memory timings" - why are you not using the default configuration? Just because it worked for months in Win7 does not mean that custom settings will be OK for Win10 too. Reverting to default settings is the first step and undoing HW changes is the next - did you try using the previous memory?

 

[sastusbulbas]

Dippy.

I am sure you do not get event 41 kernel-power simply by turning a pc off, otherwise event viewer on all 5 PC's here would have it continually. For instance currently, the PC has been switched on/off twice a day since I last made adjustments, there are NO event 41 Kernel events in events viewer since the last one on the 18th. So why make such a statement? Did I read that point wrong?

Don't wonder about memory, I thought I had clearly stated regardless of running default (load optimized defaults then boot, load optimized defaults then reboot, or enter BIOS are your options remember) or manual input of memory setting it made no difference as the random issue re-occurred, I even stated memory passed Memtest even with its XMP settings.

Not sure what you see as custom with a CPU at stock settings with memory running its specified speed timings and voltage?

No point assuming HW issue, others with this issue assumed or were given assumed hardware issues and wasted a lot of time bought new hardware and got no where. It may have helped had you read the thread and Googled event 41 kernel.

The PC has been used every day since it last happened and no issues, emphasis is being made on powering the PC on and off a couple of times a day to try and get the problem to occur and it has not, but the following changes were made at my previous post.

Advanced Power Options.
Hard Disk, Turn off changed from 20 minutes to never.
Sleep, sleep after changed to never.

Define Power Button options.
Turn on fast start up, has been unchecked.
Hibernate, this was unchecked already, though I read this must be enabled if you have fast start up enabled?

I have also read that the default settings can interfere with automatic updates as updates require a full shut down? So why have them? Was this the issue? I do remember Load optimised defaults then boot didn't work that I had to use the reboot or bios option. Could that indicate a non waking SSD due to hard drive sleep after 20 minutes or an update problem due to not fully shutting down and requiring a proper reboot that default settings was not allowing hence the hang up?

Bios.
Dummy Load enabled. (Previously set to Auto so I doubt this mattered)

I probably went a little far, as I should really have made one change at a time, I simply chose those as based on a number of articles on Google that had near identical procedures that stated those were what to try 1st.
I guess I should unwind those changes one or two at a time and see where the problem returns.

Of course if there is reason to,
I can still change power supply.
Or I can change Hard drives.
Or I can change OS.
Or I can change memory.
Or create a new Windows 10 install.

 

[Cyber-Mav]

run memory tester

 

[BUDFORCE]

I'd putting it all back, then putting back your changes, one step at a time, whatever order you did the PSU/WIN10/other thing.

Pit it all back, test it, if the problem still occurs at least you know it's not those, but assuming it is stable.

Then do one thing at a time, eg install win 10, run it for a couple of weeks, if stable, change the PSU, run for a few weeks etc.

 

[sastusbulbas]

run memory tester

Yawn.

 

[therealdippy]

Dippy.

I am sure you do not get event 41 kernel-power simply by turning a pc off, otherwise event viewer on all 5 PC's here would have it continually. For instance currently, the PC has been switched on/off twice a day since I last made adjustments, there are NO event 41 Kernel events in events viewer since the last one on the 18th. So why make such a statement? Did I read that point wrong?

Event ID 41 The system has rebooted without cleanly shutting down first - Windows Client

Describes the circumstances that cause a computer to generate Event ID 41, and provides guidance for troubleshooting the issue.

docs.microsoft.com

I believe that what you are doing is highly unusual. My expectation is that most people do what I do which is to shut down our PCs via the Windows OS. But if I have a system hang then I am forced to power off a PC - by removing the power instantaneously, nearly always by holding down the power button. In such cases I always get event 41 which is what I expect because that's exactly what Microsoft have programmed Windows to do. If you are regularly powering off your PCs without shutting down and are not getting event 41 then there has to be something strange about your installations.

 

[Tetras]

Some options I am thinking of as next steps, re-install all drivers, put his old SSD's with W7 back in and see if it re-occurs, try a fresh clean W10 install with just the basics installed on a cleared CMOS, or change the PSU. Was hoping it was an issue someone else may have had experience with. Right now after looking at the internet stories I feel its a W10 issue.

It is possible, e.g. if there's some kind of serious driver malfunction, but an OS doesn't usually cause the whole system to cut out randomly. I'd have expected it to BSOD (like the IRQL less than equal type errors) before it crashed, if that was the case.

Is there no more information on the boot failure and exactly to what it refers? It could be something as simple as a loose cable on the boot SSD, if the SSD is dropping out of the system.

 

[DK83]

is it possible you have multiple installations of windows on different drives? Ie boot manager is trying to load up a now defunct windows 7 on an older drive instead of windows 10?
maybe worth a sniff about in disk manager, make sure you only have one OS drive.

 

[sastusbulbas]

Both drives within the machine were new unused out a retail packet with a fresh USB install of Windows 10 on the 240gb drive.

The above changes to sleep/hibernate within W10 power settings removed the problem.

Now yesterday as an experiment I enabled fast start and hibernate again.

When my lad used the PC after that the crash returned.

So is that hardware related? Sadly despite plenty of reports regarding the issue, no clear answer. But its a step closer to understanding. And one step at a time.


Regarding the "Yawn" remark, well clearly previous posts state it has been tested from default to xmp and despite memory passing Memtest the problem persisted. So why say it.

Dippy, nothing I am doing is unusual, I build PC's often, and in all my years never experienced this. Yet Google it and there are plenty of others who have with no clear one fix for all.

What a joke.

 

[Chris Beard]

@sastusbulbas I think it's jolly nice of all these strangers to help you troubleshoot your problem. For free. At your request.

This video may be interesting to you: LTT video on power states

The video is about a Threadripper system, but I think there are similarities to the situation you describe.