Sudden Reboot

[Tyrannosaur]

These last few months my Windows 10 Pro PC reboots, it just loses power then automatically starts up again. No BSOD. In the past month it has happened 10 times. Can happen when it has been on for a while or not long after startup, when watching twitch or writing an email in chrome. It started a few months ago around the time I installed new drivers for my Wacom tablet. So after a few crashes the first thing I did was uninstall the drivers. The crashes continued. Looked at the Windows Reliability and no consistent error messages happen before the "previous system shutdown was unexpected". Can't see any dump files being created and Who Crashed doesn't find any either. Temperatures for all components is low.

Decided to do a fresh install of Windows 10 last week on a brand new SSD with just the basic programs I need. Thought all was well until last night when it crashed in exactly the same way.

So I'm guessing it is more hardware related?

Anything I should look at before I start changing more hardware, starting with a new PSU?

System Details:

CPU: AMD A8-5600k (not overclocked)
MB: ASUS F2A55-M
GPU: Using onboard graphics
Memory: Corsair 4x4 GB DDR3 1600
PSU: Corsair CX 430
SSD: WD Green 120 GB (was Crucial 240 GB)
HDD: 1 TB Samsung and 3 TB WD Green
USB: External HDD(powered) + Mouse/Keyboard
OS: Windows 10 Pro

Thanks for any help.

[spacein_vader]

These last few months my Windows 10 Pro PC reboots, it just loses power then automatically starts up again. No BSOD. In the past month it has happened 10 times. Can happen when it has been on for a while or not long after startup, when watching twitch or writing an email in chrome. It started a few months ago around the time I installed new drivers for my Wacom tablet. So after a few crashes the first thing I did was uninstall the drivers. The crashes continued. Looked at the Windows Reliability and no consistent error messages happen before the "previous system shutdown was unexpected". Can't see any dump files being created and Who Crashed doesn't find any either. Temperatures for all components is low.

Decided to do a fresh install of Windows 10 last week on a brand new SSD with just the basic programs I need. Thought all was well until last night when it crashed in exactly the same way.

So I'm guessing it is more hardware related?

Anything I should look at before I start changing more hardware, starting with a new PSU?

System Details:

CPU: AMD A8-5600k (not overclocked)
MB: ASUS F2A55-M
GPU: Using onboard graphics
Memory: Corsair 4x4 GB DDR3 1600
PSU: Corsair CX 430
SSD: WD Green 120 GB (was Crucial 240 GB)
HDD: 1 TB Samsung and 3 TB WD Green
USB: External HDD(powered) + Mouse/Keyboard
OS: Windows 10 Pro

Thanks for any help.

My first guess would be PSU. Do you have a spare or someone that would lend you one for a bit to see if that solves it?

[DanceswithUnix]

My first guess would also be PSU.

Also worth running memtest86 overnight.

[Tyrannosaur]

Thanks for the replies.

I do have an old spare PSU that I can try. It's a bit loud though so will see how long I can stand it.

But will run memtest86 overnight like DanceswithUnix suggests first.

[Tyrannosaur]

Ran memtest86 overnight and it did 3 passes with zero errors.

Last night I did get another crash, this time with a Blue Screen: DPC_WATCHDOG_VIOLATION. But I think I'm going to put this down to the sudden reboots messing Windows up a little.

[Tyrannosaur]

Update:

On the 17.04 changed the PSU to an old CoolerMaster 550W and last night 22.04 it once again lost power and reset. No BSOD or dump file created. One thing that did happen after this recent reset was the audio icon on the taskbar had an error saying The Audio Service Is Not Running. Although I could still get audio.

Ran the AMD Overdrive Stability test for an hour today and nothing untoward happened.

Jiggled the power switch/reset cables while the PC is running and nothing happened.

According to CPUID HWMonitor Temps are:
CPU: 35 °C
Board: 31 °C
Package: 61 °C

So any advice on what I could try next?

[Lanky123]

I seem to recall the Spectre/Meltdown patches caused this kind of issue for some people. Pretty sure they then pushed improved patches but if this started around January sort of time then it might be related?

[Strawb77]

i would clear the cmos, your mobo manual will tell you how to do it- it`s my go-to.

[Tyrannosaur]

OK reset CMOS last night. Will see how it goes.

[Tyrannosaur]

Update:

So nearly a week since I reset CMOS it once again did a power down reboot. So that didn't work.

Anything else I can try before looking for new parts?

[Saracen]

One thing I don't remember anybody suggesting, and to be honest it's the FIRST tjing that crossed my mind, is power. Not power supply, but the supply of power.

This could take many forms, from momentary powrr cuts to mains spikes. It's even possible it could be some faulty item on your home's mains, or even a neighbour. As for what, it's hard to tell but a spike from a fsulty boiler, or a freezer compressor, or even a fault in a local substation.

Your PC should be able to tolerate a reasonanable range in supply voltage, but if, for example, the PSU can cope with 230-250v, and your local supply is just nominally in that, then it doesn't take much to trigger an even.

Solution? UPS, I guess.

Or, of course, it's nothing to do with this and you have a PC fzult. But diagnostic logic suggests, at least after eliminating common problems, to start at the beginning of the chain and eliminate stages one by one.

Can you borrow an UPS?

[DanceswithUnix]

One thing I don't remember anybody suggesting, and to be honest it's the FIRST tjing that crossed my mind, is power. Not power supply, but the supply of power.

This could take many forms, from momentary powrr cuts to mains spikes. It's even possible it could be some faulty item on your home's mains, or even a neighbour. As for what, it's hard to tell but a spike from a fsulty boiler, or a freezer compressor, or even a fault in a local substation.

Your PC should be able to tolerate a reasonanable range in supply voltage, but if, for example, the PSU can cope with 230-250v, and your local supply is just nominally in that, then it doesn't take much to trigger an even.

Solution? UPS, I guess.

Or, of course, it's nothing to do with this and you have a PC fzult. But diagnostic logic suggests, at least after eliminating common problems, to start at the beginning of the chain and eliminate stages one by one.

Can you borrow an UPS?

A decent PSU should be able to ride out spikes, brownouts and even momentary power loss. I've been in the house where all the lights flickered right out and the PCs just kept going. Most of the time machines are pretty much idle, pulling maybe 50W from a 500W psu so there is bags of energy stored in the reservoir caps.

[Tyrannosaur]

Thanks Saracen for the suggestion, and I agree it would be nice to rule out the actual power supplied but unfortunately I do not have access to a UPS.

A few years ago with the same PC I did have a few ASUS startup messages about power supply surges but that has not happened with this current problem.