Pc switching off (after an hour gaming period.)

[Ellie_Tay]

----
Hello Everyone,

I am looking for some advice after having some issues with seemingly random system power offs.

What I've worked out so far:
The PC switches off while under what I will describe as moderate stress from basic gaming. Games I've experienced this include Final fantasy 14, Sea of thieves and Alien Isolation. Not that that should be important, just some idea.
Last power out resulted in 3 system power outs in a matter of 10 seconds. After being left off for a period, run as normal.


Side Notes:
I have no overclocking on this PC to my understanding, performance has been left as it was at purchase.
GPU is 5 years old. I will be replacing this part tomorrow. (20th July)
System temperatures readings seem normal using the 'HWinfo' app. (All readings under 80 Degrees)

Actions I've taken:
-Dusted the whole machine
-Moved tower to an open area
-Replaced the PSU
-Ensured components are secure & well connected
-Replaced thermal paste for both CPU and GPU
-Added additional fans ( Two exhaust fans and three intake. (Two intake at the front, one at the top. One exhaust at the top and back)
-Reconnected the front panel power cables (power button etc) ensuring properly fitted


If there is any further information you require to have a better understanding and insight, please let me know. I will be updating this post tomorrow after installing the new GPU. Though I suspect it may not be the issue.



Any help or advice or suggestions on this matter would be greatly appreciated.

Thank you,
Ellie


-----------------------------------
PC Specs


Case
LIAN LI LANCOOL II GAMING CASE
Processor (CPU)
AMD Ryzen 5 3600XT Six Core CPU (3.8GHz/35MB CACHE/AM4)
Motherboard
ASUS® ROG STRIX X570-F GAMING (USB 3.2 Gen 2, PCIe 4.0) - ARGB Ready!
Memory (RAM)
2 X 8GB Corsair VENGEANCE RGB
Graphics Card
Gigabyte GTX 960

M.2 SSD Drive
512GB PCS PCIe M.2 SSD (2000 MB/R, 1100 MB/W)

SSD
1X Sandisk 500GB SSD

HDD
1X Seagate Barracuda 2TB Internal Hard Drive HDD


Power Supply
CORSAIR 750W RMx SERIES™ MODULAR 80 PLUS® GOLD, ULTRA QUIET
Processor Cooling
STANDARD AMD CPU COOLER
Thermal Paste
STANDARD THERMAL PASTE FOR SUFFICIENT COOLING (Replaced with MX-5 Thermal Compound)
Sound Card
ONBOARD 6 CHANNEL (5.1) HIGH DEF AUDIO (AS STANDARD)
Wireless Network Card
WIRELESS INTEL® Wi-Fi 6 AX200 2,400Mbps/5GHz, 300Mbps/2.4GHz PCI-E CARD + BT 5.0
USB/Thunderbolt Options
MIN. 2 x USB 3.0 & 2 x USB 2.0 PORTS @ BACK PANEL + MIN. 2 FRONT PORTS
Operating System
Windows 10
Monitor
IIYAMA G-MASTER GB2760QSU-B1 27" (2560x1440 Resolution)

---------------------------------

 

[SpyderTracks]

Desktop graphics cards don’t take thermal paste, they require thermal pads.

Have you monitored temps? My guess is the GPU is overheating.

 

[Ellie_Tay]

Desktop graphics cards don’t take thermal paste, they require thermal pads.

Have you monitored temps? My guess is the GPU is overheating.

Not to sound argumentative or stubborn, I believe, through research, GPUs do use thermal paste as I have cleaned it off myself and reapplied following tutorials and doing some reading. Perhaps thermal pads are an alternative or there is some sort of misunderstanding.

As mentioned above, temps are below 80 degrees across the board, GPU readings are peaking at around 69 degrees, to my understanding, from the readings.

 

[SpyderTracks]

GPUs do use thermal paste as I have cleaned it off myself and reapplied following tutorials and doing some reading.

No desktop gpu uses thermal paste, they use thermal pads which are very different and have a thickness of at least 1mm.

Thermal paste wouldn’t bridge that gap.

what are you using to stress the GPU? You need to use something like the heaven benchmark or fur mark and take readings with HWMONITOR

 

[Steveyg]

No desktop gpu uses thermal paste, they use thermal pads which are very different and have a thickness of at least 1mm.

Thermal paste wouldn’t bridge that gap.

Some of the older cards do have thermal paste that you can reapply. They've been superseeded by Thermal Pads these days but some of the older ones still have the paste like the below video

 

[Scott]

3 power outages in 10 seconds? Do you mean that the system rebooted 3 times within 10 seconds or physically switched off?

The PC switching off completely without rebooting tends to be temperature related, in my experience the CPU.

Anything regarding the GPU tends to crash to desktop but there are certain power situations where it will reboot or shut off.

From your testing, and presuming all your findings are as I take them, I think the GPU is the next logical step. Having said that, the first thing I would have checked would have been the memory so I would definitely run a memtest on the rig....... this would be more aimed at you experience reboots rather than hard shutdowns though.

 

[Ellie_Tay]

3 power outages in 10 seconds? Do you mean that the system rebooted 3 times within 10 seconds or physically switched off?

The PC switching off completely without rebooting tends to be temperature related, in my experience the CPU.

Anything regarding the GPU tends to crash to desktop but there are certain power situations where it will reboot or shut off.

From your testing, and presuming all your findings are as I take them, I think the GPU is the next logical step. Having said that, the first thing I would have checked would have been the memory so I would definitely run a memtest on the rig....... this would be more aimed at you experience reboots rather than hard shutdowns though.

Thank you Scott, I will look into this, it is some very helpful insight. I am more than happy to give this a try in order to attempt to pinpoint the issue.

In relation to your question, I described it as a 'power off' because the system seemed to lose all power and reboot from the beginning. All fans slowed as though no power was being delivered between the time it seemed to switch off and the boot up sequence.

Thanks again for your response.

 

[Scott]

Thank you Scott, I will look into this, it is some very helpful insight. I am more than happy to give this a try in order to attempt to pinpoint the issue.

In relation to your question, I described it as a 'power off' because the system seemed to lose all power and reboot from the beginning. All fans slowed as though no power was being delivered between the time it seemed to switch off and the boot up sequence.

Thanks again for your response.

This is good and bad news. It's good as it's not a hard shutdown that's being caused...... it's bad because it's effectively a non-BSOD reboot..... which can be caused by a number of things.

RAM & GPU would definitely be my 2 focal points for now. You're replacing the GPU tomorrow anyway, so see how that goes, and then look to run 1 stick of RAM to see how that goes.

As long as you can create the issue you should have no problem finding out what it is that's causing it.

I would be curious to see what the event viewer holds after a reboot.

 

[Fabian03]

No desktop gpu uses thermal paste, they use thermal pads which are very different and have a thickness of at least 1mm.

Thermal paste wouldn’t bridge that gap.

what are you using to stress the GPU? You need to use something like the heaven benchmark or fur mark and take readings with HWMONITOR

The gpu die itself paste is used then depending on aib vrm and other things are bridged my pads or thermal putty in the case of some evga cards.

 

[Ellie_Tay]

This is good and bad news. It's good as it's not a hard shutdown that's being caused...... it's bad because it's effectively a non-BSOD reboot..... which can be caused by a number of things.

RAM & GPU would definitely be my 2 focal points for now. You're replacing the GPU tomorrow anyway, so see how that goes, and then look to run 1 stick of RAM to see how that goes.

As long as you can create the issue you should have no problem finding out what it is that's causing it.

I would be curious to see what the event viewer holds after a reboot.

I've just completed the memtest, 3 hours is a lengthy time.
I used memtest 86 for that. It's come back with no errors from four runs of the test with the two ram sticks inserted.

I also downloaded AIDA 64 and run stress tests on CPU,GPU and stogage for fifteen minutes, no crash. Perhaps it's a more timely event that triggers the problem like a build up of heat.


I'll check out the event viewer when it occurs again, though I'm not very familiar myself, some reading first perhaps.
I guess as you say, I can cross out gpu tomorrow at least.

Thanks

 

[ubuysa]

I'll check out the event viewer when it occurs again, though I'm not very familiar myself, some reading first perhaps.
I guess as you say, I can cross out gpu tomorrow at least.

See this post: https://www.pcspecialist.co.uk/forums/threads/when-youre-seeking-help-with-a-bsod.71885/#post-568901. Although it talks about BSODs there are instructions in there for uploading your System and Application logs. There may well be something in there that points at your problem.

 

[Ellie_Tay]

This is good and bad news. It's good as it's not a hard shutdown that's being caused...... it's bad because it's effectively a non-BSOD reboot..... which can be caused by a number of things.

RAM & GPU would definitely be my 2 focal points for now. You're replacing the GPU tomorrow anyway, so see how that goes, and then look to run 1 stick of RAM to see how that goes.

As long as you can create the issue you should have no problem finding out what it is that's causing it.

I would be curious to see what the event viewer holds after a reboot.

Okay so after installing the new GPU, closing the case and running a game for 20 minutes, the problem has still occurred.
Again, pc has turned off and attempted rebooting a few times before I turned it off fully myself.

As I said before Memtest 86 was clear of errors during it's testing for the RAM.

I'm honestly baffled, I've no other ideas personally.
& I'm unsure what exactly I'm looking for in the event viewer. Unsure if I should be allowing people to view that either. Doesn't make sense to share data when you don't know the potentially stored information of it


Small edit/update: I did find something in the event logs. Could be something, could be nothing:
The device HID-compliant headset (location (unknown)) is offline due to a user-mode driver crash. Windows will attempt to restart the device 5 more times. Please contact the device manufacturer for more information about this problem.

 

[B Poot]

GPU is 5 years old.

Out of interest, have you done a clean install of Windows in the past year or so?

Just trying to cover all bases as it seems like you have done some good checks on the hardware. Perhaps if you do a clean install of Windows, get all the updates and then make sure your hardware drivers are up to date, that might help (it certainly wouldn't hurt).

 

[Scott]

Okay so after installing the new GPU, closing the case and running a game for 20 minutes, the problem has still occurred.
Again, pc has turned off and attempted rebooting a few times before I turned it off fully myself.

As I said before Memtest 86 was clear of errors during it's testing for the RAM.

I'm honestly baffled, I've no other ideas personally.
& I'm unsure what exactly I'm looking for in the event viewer. Unsure if I should be allowing people to view that either. Doesn't make sense to share data when you don't know the potentially stored information of it


Small edit/update: I did find something in the event logs. Could be something, could be nothing:
The device HID-compliant headset (location (unknown)) is offline due to a user-mode driver crash. Windows will attempt to restart the device 5 more times. Please contact the device manufacturer for more information about this problem.

Do you have a headset connected?

Going forward with the troubleshooting remove everything from the system other than critical peripherals to carry out the tests.

It could be something like a faulty USB port that's causing the problem, or the headset itself, or something connected.

For the base point for testing you want minimum inputs.

 

[ubuysa]

Okay so after installing the new GPU, closing the case and running a game for 20 minutes, the problem has still occurred.
Again, pc has turned off and attempted rebooting a few times before I turned it off fully myself.

It's probably not the GPU then.

As I said before Memtest 86 was clear of errors during it's testing for the RAM.

Memtest can only prove that your RAM is bad, it can never prove that your RAM is good. It must still be a suspect.

I'm honestly baffled, I've no other ideas personally.
& I'm unsure what exactly I'm looking for in the event viewer. Unsure if I should be allowing people to view that either. Doesn't make sense to share data when you don't know the potentially stored information of it

It's up to you of course but the event logs were designed to be viewed by a third party. There is nothing in them that is either personal nor security related. It's your call though.

Small edit/update: I did find something in the event logs. Could be something, could be nothing:
The device HID-compliant headset (location (unknown)) is offline due to a user-mode driver crash. Windows will attempt to restart the device 5 more times. Please contact the device manufacturer for more information about this problem.

You'll find lots of errors like this in the event logs. Everything that fails, and a lot of stuff that succeeds, is written to the log. It's impossible to say whether that is significant on its own.

 

[Ellie_Tay]

Do you have a headset connected?

Going forward with the troubleshooting remove everything from the system other than critical peripherals to carry out the tests.

It could be something like a faulty USB port that's causing the problem, or the headset itself, or something connected.

For the base point for testing you want minimum inputs.

It's probably not the GPU then.

Memtest can only prove that your RAM is bad, it can never prove that your RAM is good. It must still be a suspect.

It's up to you of course but the event logs were designed to be viewed by a third party. There is nothing in them that is either personal nor security related. It's your call though.

You'll find lots of errors like this in the event logs. Everything that fails, and a lot of stuff that succeeds, is written to the log. It's impossible to say whether that is significant on its own.

Okay so, in response to both of your posts, I have taken out the headset from the USB slot and uninstalled the " HID-compliant headset" drivers.
Following this, the system has turned off once again.
I'm still uncertain on what information are in event logs, however, decided to put some faith in this helpful community. I've not got any cloud stoages set up at the moment, any ideas on the easiest to use would be great.

From my perspective it seems as though something is overheating. Though the highest temperature readings I have are around 67 Degrees.

 

[Scott]

Try HWMonitor for the temperatures. Just make sure you have all the options opened up to view it.

Over-temp tends to shut the PC off altogether though, rather than reboot.

 

[Ellie_Tay]

Try HWMonitor for the temperatures. Just make sure you have all the options opened up to view it.

Over-temp tends to shut the PC off altogether though, rather than reboot.

I've been using both HW Monitor and Hwinfo for temp checks

Edit:
I've added pictures of my normal temperatures, no stress.
There are two questionable temperatures, but I believe they are false readings. Even if not, I'm unsure of their purpose or location on the board.

 

[ubuysa]

Okay so, in response to both of your posts, I have taken out the headset from the USB slot and uninstalled the " HID-compliant headset" drivers.
Following this, the system has turned off once again.
I'm still uncertain on what information are in event logs, however, decided to put some faith in this helpful community. I've not got any cloud stoages set up at the moment, any ideas on the easiest to use would be great.

From my perspective it seems as though something is overheating. Though the highest temperature readings I have are around 67 Degrees.

I understand your caution. If you're not comfortable sharing the event logs then don't do it. We'll troubleshoot your system in other ways.

 

[Ellie_Tay]

I understand your caution. If you're not comfortable sharing the event logs then don't do it. We'll troubleshoot your system in other ways.

Thank you for your understanding, I have edited the comment above the comment I'm replying to with an image displaying temperatures. Your thoughts would be greatly appreciated.
Also, do you have a recommended cloud service? I did try dropbox but it seems to be difficult without recipients emails etc.