Random Shutdowns

I've been having a very strange issue with random shutdowns lately. Sometimes I get them at idle, sometimes when gaming/benchmarking. I can can go days at a time without it happening though. Weird thing is, it doesn't just shutdown, the power led on the front of the laptop starts blinking orange and what I believe to be the gpu fan starts to spin at high speed. It does this anywhere for about 15-30 seconds before powering off. I can still use the laptop normally during the time before it shuts down, so if I'm playing a game I can usually finish what I'm doing and save my progress really quick before the 'timer' ends. GPU and CPU temps are good, both are usually below 70C when this happens, and as I've said before it sometimes happens at idle. I've never had any problems turning it right back on.

Weirdest thing is, before the laptop turns off, it will show the 'Windows is shutting down' screen for a moment, almost as if it is trying to go through a proper shutdown sequence. So something seems to be telling Windows to shut itself down, which at first glance would seem to be a software issue, however the flashing power led and fans point to some kind of hardware fault. I've tried just about all the troubleshooting in the book, including removing CMOS battery, reinstalling Windows, reverting to Windows 7/8.1 (currently using 10), different gpu drivers, memory test, even measuring power supply wattage using a kilowatt monitor and trying different outlets. There is also nothing in the event viewer except a generic critical error stating that the computer did not shutdown properly.

I've also let it sit in the BIOS for several hours to see what would happen, but no shutdown occurred. Ran a live usb version Lubuntu for a day as well with no issues. It's a strange problem for sure, and I'm still trying to figure it out before having to RMA, but it's looking like I eventually will have to. The most telling sign that something is seriously wrong is the blinking power indicator and fans, but I have no idea what this could indicate. Perhaps someone can chime in here with some insight on this or point me to something that I may have missed.

 

The fact it didn't do it with Lubuntu makes me wonder if it isn't a graphics driver issue. Lubuntu comes with open source drivers only. The wide range of hardware support should eliminate most other things in the machine as the culprit. I would try installing the proprietary drivers and see if the problem crops up but I would use Ubuntu since it uses desktop composition while Lubuntu does not last time I ran it.

 

I was thinking that these symptoms seemed 'out of scope', so to speak, for a driver. I could see drivers causing a blue screen or restart, but blinking LEDs and ramping up the fans seems kind of odd. Nevertheless, I am definetely open to suggestions and willing to exhaust all possibilities. The problem has been occurring with drivers ranging back to 359.00 (currently using latest 361.43), and I haven't seen any reports on the forum here or anywhere else with anyone having driver issues with the 980m, but the intermittent nature makes it a little harder to diagnose. What I'm going to try first is using DDU and reverting back to the latest gpu driver posted by Sager themselves to see if that makes a difference.

As far as trying Ubuntu I'll have to do that as well. Main reason I chose Lubuntu was because I only have a 1gb USB stick, and I need at least 2gb for Ubuntu, but I'll just purchase another one and run it for awhile to see what happens since you say Ubuntu uses the gpu more. Not having at least a 32gb USB stick in 2016 is kind of ridiculous anyways lol.

 

I've had issues with 359 and up with my 980Ms... The 361.43 actually causes a hard shutdown and restart every time I try to install them. I would use DDU to remove the drivers and install 347.88, its the most stable driver for the 980M that I've used and most agree that 347.88 is incredibly stable.

 

The lights flashing and fans revving up is an EC (electronic control) of the machine going into a panic. Some value it monitors be it temperature, voltage or fan speed is out of range (due to either the value or sensor having an issue) and the EC is shutting down the machine to prevent damage.

 

@Ethrem Currently running the stock Sager driver to see if that makes a difference.

@Meaker The EC explanation intrigues me and is the info I've been looking for as it potentially solves a part of the puzzle. I do undervolt the CPU using XTU, but I figured this could not be causing my problem due to the fact that every system (including this one) in which I've set the cpu voltage too low has just resulted in a simple reboot, not an EC panic and shutdown. It does seem less likely to happen the lower I have the voltage offset, but it could just be a coincidence with the the issue being so random and sporadic in nature. I doubt it's a temp sensor as I have logging enabled with HWiNFO and it has never reported anything higher than 75C for the cpu or 70C for the gpu, unless maybe there is some other sensor involved that I have been ignoring. But since you did mention voltage, is it possible that undervolting is causing this?

 

Usually not, I would double check all the fans are spinning properly first and they are securely seated.

 

Kind of a necro but I wanted to just update this thread (I hate researching a problem and finding a forum thread somewhere that's related in which the OP just vanishes without a resolution). Ultimately I ended up having to RMA and have the motherboard replaced. Of course Sager wouldn't give me any more information than that (Had to email them to even find out that the board was replaced since there is minimal communication and no other documentation sent back with the machine). As to what specifically was wrong with the board, they probably don't know themselves other than it was 'bad'. So if anyone else experiences this problem you may need a new board. Not much of a resolution but there you have it lol.

 

If a board fails any testing it will be replaced, we may to further testing down the line to determine exactly what went wrong to keep an eye on any patterns but that wont be done on your time

 

@Meaker

Yea that makes sense. The RMA was handled nice and fast which is much appreciated, and I haven't noticed any other issues. I'm just the kind of person that likes to know why something went wrong and what specifically caused it. But like you say it's probably better for the end user to just replace the part and call it a day since no one likes to be without their $2000+ laptop for very long.