Computer keeps restarting

[Kristjan]

I made an account on this website just to ask for a bit of help. Pretty much only know the basics about a computer.

For a few weeks my computers been restarting randomly. Started off with once or twice a day and currently restarts every few minutes. I've noticed however that sometimes it's able to go for up to an hour before restarting, very rarely tho. It only restarts when windows is running. While I was reinstalling it, it ran without a problem. Same thing when updating the bios.
I've tried to find a way to fix it myself, but can't find the correct solution.

I've tried getting a new psu and switched the power as some forum posts lead me to believe it might be that, but it didn't work. I've also tried to completely reinstall windows as well as update all my drivers along with the bios. I've also tried taking apart my computer and cleaning the parts separately in case it might help.

The only things I can think of now are ram processor or motherboard, but I don't know how I'd go about checking which one is the problem if even any. Been trying to find a solution for a good while now and I'd appreciate any suggestions.

Also here are the event logs. I can't make sense of them was hoping someone could tho. Karnel eror appeared everytime the computer rebooted.

 

[looniam]

welcome,

pretty brave to update the bios when it could shut down.

more system specs will help esp. brand model of PSU, motherboard, and then cpu (Oc/stock?) ram and graphics card. (fill out system specs in your profile please)

 

[Kristjan]

welcome,

pretty brave to update the bios when it could shut down.

more system specs will help esp. brand model of PSU, motherboard, and then cpu (Oc/stock?) ram and graphics card. (fill out system specs in your profile please)

Was honestly pretty sure it must be a bios problem so I took the risk

Specs:
new PSU: argus APS 620W
old PSU: ATX LC420-12 V2.31 (I think that's it, it says it on the power supply)

Motherboard: asrock h81m-dg4
CPU: intel i5 4460
RAM: 8gb
GPU: GTX 1050 ti

 

[looniam]

Was honestly pretty sure it must be a bios problem so I took the risk

Specs:
new PSU: argus APS 620W
old PSU: ATX LC420-12 V2.31 (I think that's it, it says it on the power supply)

Motherboard: asrock h81m-dg4
CPU: intel i5 4460
RAM: 8gb
GPU: GTX 1050 ti

oh my, did you spend ~30 euro for that no name PSU? really need to spend 2x that much for +/< 550 watts.

everything else is rather simple; 250 watts can power all that including a few fans and HDs.

edit:
though if you want to try something, does the 10950ti have a 6 pin power? sort of no matter but if you have a monitor to connect DVI/VGA on the motherboard (set bios to igpu) and pull the card to see if any changes.

but the PSU screams at me.

 

[pavle]

Try another/third PSU, if you can.

 

[Kristjan]

oh my, did you spend ~30 euro for that no name PSU? really need to spend 2x that much for +/< 550 watts.

everything else is rather simple; 250 watts can power all that including a few fans and HDs.

Yeaa it was a buy of an impulse buy honestly. I thought it'd be a good idea to test out if the psu was really the problem.
So the psu I bought might not be providing enough power?
And yeah the gpu does have 6 pin power I'll try that, but what do you mean by set bios to igpu?

 

[Shrek]

Specs:
new PSU: argus APS 620W
old PSU: ATX LC420-12 V2.31

Same problem on both supplies?

 

[Kristjan]

Same problem on both supplies?

Yea same thing happens. Previous posts made me sceptical about the new supply tho since it's an off brand.

Same problem on both supplies?

I'm running the new psu without the graphics card right now, waiting to see if it reboots.

 

[looniam]

Yeaa it was a buy of an impulse buy honestly. I thought it'd be a good idea to test out if the psu was really the problem.
So the psu I bought might not be providing enough power?
And yeah the gpu does have 6 pin power I'll try that, but what do you mean by set bios to igpu?

well, it might be throwing good money at bad seeing a "no-name" unit like that and the low price. its not about having enough power but a flaw in parts/manufacturing.

come to think of it, your bios settings for graphics might be auto but i am making sure you have a display out when you pull the card and reboot the system. i don't know what other parts you can cannot swap out for trial and error troubleshooting.

 

[Shrek]

I'd follow with a thorough RAM test

 

[Kristjan]

well, it might be throwing good money at bad seeing a "no-name" unit like that and the low price. its not about having enough power but a flaw in parts/manufacturing.

come to think of it, your bios settings for graphics might be auto but i am making sure you have a display out when you pull the card and reboot the system. i don't know what other parts you can cannot swap out for trial and error troubleshooting.

I never thought the psu could still be the problem since it's new. I took out my gpu and so far it's been running for about 30 minutes which is rare. That's more of a red flag for the psu than the gpu right?

Yea I don't remember ever changing my bios at all so it should be default. And I sadly have no spare parts to swap out and test.

I'd follow with a thorough RAM test

Is there a way to test it without having a second ram stick?

 

[Shrek]

Run Windows Memory Diagnostics

and/or

MemTest

 

[Deleted member 24505]

If this argus APS 620W is your PSU it is a turd, change it imo

 

[looniam]

I never thought the psu could still be the problem since it's new. I took out my gpu and so far it's been running for about 30 minutes which is rare. That's more of a red flag for the psu than the gpu right?

naw, ya ain't out of the woods yet.

try running something ultra light like CPU-Z benchmark, then try cinebench for something a little heavier.

fwiw, i just recalled a few years ago i went through random shut downs after 2 known good PSUs - long story short - i believe there was a loose screw behind my motherboard.
i don't keep anything "spare" in my case no more.

 

[Kristjan]

naw, ya ain't out of the woods yet.

try running something ultra light like CPU-Z benchmark, then try cinebench for something a little heavier.

fwiw, i just recalled a few years ago i went through random shut downs after 2 known good PSUs - long story short - i believe there was a loose screw behind my motherboard.
i don't keep anything "spare" in my case no more.

Seems I accidentally deleted my last reply.
I did a test on cpu z and cinebench and it seems good. This would mean the gpu isn't getting enough power from the psu right? If so what should I be on the lookout for when buying a new one seeing as I messed up this time?

Oh and as I said in the comment I deleted I really appreciate all the help, I'd be pretty lost right now.

 

[DeathtoGnomes]

I'd like to see the windows event viewer errors, specifically the big red ones.

Also, if you let windows auto update, it will restart your pc constantly after a fresh install, and sometimes without warning.

 

[looniam]

Seems I accidentally deleted my last reply.
I did a test on cpu z and cinebench and it seems good. This would mean the gpu isn't getting enough power from the psu right? If so what should I be on the lookout for when buying a new one seeing as I messed up this time?

Oh and as I said in the comment I deleted I really appreciate all the help, I'd be pretty lost right now.

to know for sure is trying out the 1050ti in another system.
really both cpu-z and cinebench are fine more for "stable enough" than fully stressing the PSU. aida64 extreme stability test(s) pushes it a little farther PSU wise. it would be great if you know someone to swap GPUs real quick.

you're functional now, i'd still consider breadboarding. though those two PSUs you listed are questionable, even w/the graphics card, your system isn't needing 250 watts total. those are some rotten potatoes then . . no offence.

 

[RJARRRPCGP]

The "critical" ones are often just because Windows wasn't shut down properly. (this will happen with NT 6 and later.)

What we need to see, is if you have any WHEA errors or bugcheck reports.

 

[Kristjan]

I'd like to see the windows event viewer errors, specifically the big red ones.

Also, if you let windows auto update, it will restart your pc constantly after a fresh install, and sometimes without warning.

I've disabled auto updates, but it is possible I scheduled one and forgot about it a while back. Is there a way to check/revert that?

to know for sure is trying out the 1050ti in another system.
really both cpu-z and cinebench are fine more for "stable enough" than fully stressing the PSU. aida64 extreme stability test(s) pushes it a little farther PSU wise. it would be great if you know someone to swap GPUs real quick.

you're functional now, i'd still consider breadboarding. though those two PSUs you listed are questionable, even w/the graphics card, your system isn't needing 250 watts total. those are some rotten potatoes then . . no offence.

I celebrated a bit early it seems, while trying to reply on my computer without the gpu being in it, same as yesterday it rebooted. Guess it was just a temporary fix yesterday, it's back to it's old self restarting every few minutes.
I did see a similar forum post that questioned the motherboard. Is there a way to check if that's the problem without having a spare?

The "critical" ones are often just because Windows wasn't shut down properly. (this will happen with NT 6 and later.)

What we need to see, is if you have any WHEA errors or bugcheck reports.

Those would be in the event logs right? If so I didn't see any, the two errors I got were 41 karnel power and eventlog 6008

 

[xorstl]

How are the the temps looking like? If you get 0 drivers errors/bluescreens and your memories are fine and can run cinebench but only.. sometimes? Sounds like something is overheating and mb protection kicks in. Not sure how well equipped with sensors that MB is but if you want to run whatever tasks you do that makes the PC reboot while having HWMonitor on, you might be lucky to catch a sudden temp spike right before the reboot.

Might be worth checking windows system logs too, might catch a specific driver failing (windows might be able to recover from such driver crash and reboot, rather than bluescreen). Not sure if this guide is decent but there are plenty others out there for the same goal: https://www.windowscentral.com/how-find-reason-pc-shutdown-no-reason-windows-10#:~:text=To check the Event Viewer,result to open the console.&text=Right-click the System category,the Filter Current Log option.

Edit: seems I missed a reply and someone already asked for the event logs. Neither of the errors you posted are what we're looking for. 6008 is just a consequence of the the 41, which is a sudden, unclean reboot. We're looking for the cause of that reboot

Is the PC on idle or doing something when it reboots? Could be interesting to see if the reboot happens always in the same cpu C state, which would indicate an almost certain PSU issue. Your CPU is not new so I don't expect the PSU to be lacking functionality but who knows, it's a shitty one.. The CPU is a haswell, if the PSU is pre-haswell you might have a problem You can test this easily by disabling ErP/EuP in your MB bios (please google where that is, it depends on the MB and bios version, but basically turn off anything related to "Deep sleep" or "Power Saving Mode").


Update: Just saw that your MB only needs a single 4pin ATX 12V connector. Are you connecting the correct cable there? Only asking because you also have a 4+4 EPS connector. While this is compatible with cpu 12V 8 pin, I'm not sure half of an EPS 4+4 12V is compatible with 4pin 12V (maybe someone else knows), so just checking you're not using half of the 4+4pin instead of the single 4 pin cable from the PSU. And btw yeah this PSU is not only bad/cheap but also really old, not sure if you knew this and went for it because your MB/cpu are also old, but this PSU can't power almost any modern system.

As expected with shitty PSUs the real technical specs documented by the manufacturer are nowhere to be found and the information in each site is different.

Different sites, different info. Please confirm what connectors it actually has and which ones you're using.

 

[Kristjan]

How are the the temps looking like? If you get 0 drivers errors/bluescreens and your memories are fine and can run cinebench but only.. sometimes? Sounds like something is overheating and mb protection kicks in. Not sure how well equipped with sensors that MB is but if you want to run whatever tasks you do that makes the PC reboot while having HWMonitor on, you might be lucky to catch a sudden temp spike right before the reboot.

Might be worth checking windows system logs too, might catch a specific driver failing (windows might be able to recover from such driver crash and reboot, rather than bluescreen). Not sure if this guide is decent but there are plenty others out there for the same goal: https://www.windowscentral.com/how-find-reason-pc-shutdown-no-reason-windows-10#:~:text=To check the Event Viewer,result to open the console.&text=Right-click the System category,the Filter Current Log option.

Edit: seems I missed a reply and someone already asked for the event logs. Neither of the errors you posted are what we're looking for. 6008 is just a consequence of the the 41, which is a sudden, unclean reboot. We're looking for the cause of that reboot

Is the PC on idle or doing something when it reboots? Could be interesting to see if the reboot happens always in the same cpu C state, which would indicate an almost certain PSU issue. Your CPU is not new so I don't expect the PSU to be lacking functionality but who knows, it's a shitty one.. The CPU is a haswell, if the PSU is pre-haswell you might have a problem You can test this easily by disabling ErP/EuP in your MB bios (please google where that is, it depends on the MB and bios version, but basically turn off anything related to "Deep sleep" or "Power Saving Mode").


Update: Just saw that your MB only needs a single 4pin ATX 12V connector. Are you connecting the correct cable there? Only asking because you also have a 4+4 EPS connector. While this is compatible with cpu 12V 8 pin, I'm not sure half of an EPS 4+4 12V is compatible with 4pin 12V (maybe someone else knows), so just checking you're not using half of the 4+4pin instead of the single 4 pin cable from the PSU. And btw yeah this PSU is not only bad/cheap but also really old, not sure if you knew this and went for it because your MB/cpu are also old, but this PSU can't power almost any modern system.

As expected with shitty PSUs the real technical specs documented by the manufacturer are nowhere to be found and the information in each site is different.
View attachment 233707


View attachment 233705
Different sites, different info. Please confirm what connectors it actually has and which ones you're using.

I'll test to see if the computer overheats and I'll update you on it.
The computer reboots no matter what I do on it, even tested out not having a single program running and left it on the homescreen, it rebooted within minutes.
With the bios tweak, how can I go about testing of that ends up being the problem? I'll turn deep sleep off now tho.
Yea my motherboard is connected with one of the 4 pins, although my old psu has the same double 4pin connection and it ran without an issue for round 6 years.
I'm aware my specs are pretty old I've had this computer since round 2016/2017 ish and only changed the gpu around two years ago.

So do you not think it's possible for the motherboard to be the problem? I'd prefer it if the problem lies with the psu but at this point I'm losing hope

One more thing I just noticed while in the bios, my cpu temperature seems a bit high. Did a quick search and it should be under 80 Celsius usually right? While idle in the bios mine was pretty stable at a good 94 Celsius.

Edit: it seems deep sleep is already disabled

Edit: while idle the cpu is around 75-80 Celsius but did hit a max of 100 at the start, don't know if that helps at all

 

[looniam]

i hate to say it but you're pretty stuck w/o being able to swap out parts. it a guess until you either replace whats faulty and it stops turning off or you put it in another rgis and starts acting up.

if you have no money or friends like me you can keep an eye on office machines getting tossed. aside from dell having a proprietary 24 pin (sometimes) most of those should be enough to test w/stripped down rig; mobo, cpu, ram all you need is the 24 motherboard and 4 pin for cpu. maybe think of breadboarding in the meantime. and watch some videos if that term is foreign.

 

[Kristjan]

i hate to say it but you're pretty stuck w/o being able to swap out parts. it a guess until you either replace whats faulty and it stops turning off or you put it in another rgis and starts acting up.

if you have no money or friends like me you can keep an eye on office machines getting tossed. aside from dell having a proprietary 24 pin (sometimes) most of those should be enough to test w/stripped down rig; mobo, cpu, ram all you need is the 24 motherboard and 4 pin for cpu. maybe think of breadboarding in the meantime. and watch some videos if that term is foreign.

I was hoping that wouldn't be the case. Machines don't really get tossed aside all that much here so that's a bit difficult to come by. I'm thinking of buying an identical motherboard tho and seeing if anything changes.
First time hearing that term, I'll look into it.

Thanks for all the help man

 

[xorstl]

One more thing I just noticed while in the bios, my cpu temperature seems a bit high. Did a quick search and it should be under 80 Celsius usually right? While idle in the bios mine was pretty stable at a good 94 Celsius.

Edit: it seems deep sleep is already disabled

Edit: while idle the cpu is around 75-80 Celsius but did hit a max of 100 at the start, don't know if that helps at all

No this is not normal at all and 99.9% the problem. Your CPU does little to no effort while you're in the BIOS. That generation doesn't even pull the power to justify 100ºC under extreme load in Africa.

Your cooler is most likely not correctly mounted and not making any contact with the CPU. Fix that and you'll fix your PC. You need to see idle temps of 30ºs C, not 90ºs (Ok, 40ºs if you're actually in Africa or a comparable hot place). If you're still using the thermal paste from 2014, please change it and make sure to stay way from metalic thermal pastes as you'll just damage your PC further.

Had you mentioned the temps in your first post you'd probably have the answer within 5 minutes

PS: if you're 100% sure the heatsink is correctly assembled and the thermal paste properly applied, the only other thing I can think of that justifies those temps is a short. I would assume your mb has short protection where needed and would shutdown immediately, or your PSU, that's out of my field of expertise, but some of the guys here can help with electronics! Nonetheless, it wouldn't run for 5-10min at 90ºC and THEN shutdown. This really sounds like improper heatsink assembly.

 

[outpt]

What kind of cooler? That’s some dam high cpu temperatures at idle. Maybe it’s rebooting to keep the cpu safe.