Spontaneous Reboot

[Denna]

Is there a known issue of a recent Windows 10 Update that causes spontaneous reboots on a system that didn't have them before.

After connecting the PC to the Internet to download a Steam game, I later noticed that Windows 10 had downloaded a large update - possibly Windows 1803.

Whenever I start a game like Batman Arkham City or even Windows 10 Solitaire, after about 30 seconds, the screen goes black and the PC reboots.

This doesn't happen in any other type of program like browsers, office apps, File Explorer, etc.

Updating to the latest Nvidia GPU driver 398.x doesn't correct the rebooting.

The PC was running without problems for years.

I've reseated the GPU and memory modules, tried one of the two memory modules at a time, but the reboots persist.

Sound familiar ?

[Silent_Scone]

At this stage, I would definitely consider trying both options.

[Denna]

Silent Scone wrote:
At this stage, I would definitely consider trying both options.

Silent Scone,

Updating to the NV316MH.1K1 firmware didn't fix the problem either.

[Lu_GRU]

@Denna
These information provided by Your last two msg should be in Your opening one, also You should double check data when providing them (graph temps).

I would do this:
- usb live Linux distro and testing -> if Linux works then it's Windows / drivers issue no HW one (also no need to install anything so fastest way to test "new OS")

If above fails, either:
-- reapply factory cooling solution, least expensive and if, with stock cooling, this also happens then it's much deeper problem beyond home repair I would guess
-- redo whole cooling on GPU if 1st step passes, reapply GPU paste, new thermal pads for vrm / memory <- !!! check them maybe first, depending on thermal-pads they could just go bad (4 years, on maximum stress I would assume, could wear them off)

[Denna]

I corrected the image in post #26, changing "Temp" to "Temp Limit".

https://rog.asus.com/forum/showthread.php?103898-Spontaneous-Reboot&p=731732#post731732

As for updating the vBIOS to a newer version, that occurred yesterday, not July 22 when this thread was started.

After erasing the SSD and reinstalling the previous version of WIndows 10 (Build 1703), the problem still occurs.

I'd say the problem is a non-repairable hardware issue on the video card. There doesn't appear to be any obvious damage on the video card.

There's nothing else to check.

[davemon50]

Sounds trivial, but have you blown the card clean of dust, most particularly the fans? If they stop, cycle, or slow unexpectedly during a game the temps will skyrocket. Have you noticed any system sound changes prior to unwanted shutdowns?

[Denna]

davemon50 wrote:
Sounds trivial, but have you blown the card clean of dust, most particularly the fans? If they stop, cycle, or slow unexpectedly during a game the temps will skyrocket. Have you noticed any system sound changes prior to unwanted shutdowns?

davemon50,

Years ago, when the overclock settings for the GPU were being determined, GPU stress tests would push the temperature into the 80's C range for minutes at a time with the system remaining stable.

The vBIOS, GPU driver, GPU driver settings, games or other hardware hadn't changed when the reboots started.

After stress testing the CPU and memory without reboots, I deduced the problem to be a video card hardware failure.

The video card does appear to be more stable when reflashing the manufacturer's stock vBIOS and not enabling any overclocking.

However, the GPU core now idles at 40 C whereas it use to idle at 27 C.

The reboots occur when the GPU core temperature reaches 65 C. When running a modern game on a modern video card, 65 C shouldn't cause a reboot.

I think the rebooting occurs when an application causes the video card to demand more power and the VRMs aren't able to handle load due to chip degradation caused by overclocking over an extended period of time (i.e., years).

It's just time for a new video card.

I can't tell if two years with a high overclock and running games that created heavy power demands on the GPU means MSI's GTX 970 Gaming 4G lasted a long time due to better than average quality components or it didn't last a long time due to lower than average quality components.

[Lu_GRU]

@Denna
Did You swapped stock cooling and tried after that?
If that didn't help then yeah more or less card is beyond home repair.

According to Your findings it seams like some of inductor / capacitor / mosfet could get fried and that could affect some power phases.
Here is more info:
http://www.geeks3d.com/20100504/tutorial-graphics-cards-voltage-regulator-modules-vrm-explained/

[Denna]

Lu_GRU wrote:
@Denna
Did You swapped stock cooling and tried after that?
If that didn't help then yeah more or less card is beyond home repair.

According to Your findings it seams like some of inductor / capacitor / mosfet could get fried and that could affect some power phases.
Here is more info:
http://www.geeks3d.com/20100504/tutorial-graphics-cards-voltage-regulator-modules-vrm-explained/

Lu_GRU,

I didn't reinstall the stock cooling, because it would have provided less cooling than the one installed.

Thanks for the VRM article link. I watched a few GPU VRM videos that discussed similar points.