Random Reboot Issue - Computer stops outputting signal and resets on its own while gaming / Kernel power ID 41

mgadeleo

Reputable
Apr 16, 2017
11
0
4,520
I don't know if this is the right subforum for this but here it goes. For the past two years I've been using a system that I built myself. Gaming computer. Its been working like a charm since then, until a couple of weeks ago. I was playing Rocket League when my display showed a solid color for about a second, and I hear a repeating sound, like when computers hang up and you hear the last second of audio repeatedly. The solid color and the static-like/repeating sound happen for about a second, and then the screen goes black, and then the monitor, or in this case TV, shows no input. I could still hear the audio but I had no image. All I can do at that moment is hit the reset button, although sometimes it resets on its own. It has also happened that the system reboots itself while watching Netflix, but this is instant. It doesn't include the solid color and display problem.

So it seems like there are two separate problems, or maybe just one with two instances.
Problem #1: Display goes dark for 1 second, solid color and static noise for 1 second, and then it just stops outputting anything. Happens while gaming, so far in Rocket League, League of Legends, and Insurgency. It also happened while I was editing a Word file once, which makes me think it's not a GPU problem, as Microsoft Word is hardly intensive, or not at all. I just hit the reset button, the boot device led on my motherboard comes on and computer restarts and works as usual, until the next time it crashes, or doesn't.

Problem #2: Happens while watching Netflix or Youtube videos. This is instant. No color and no audio issue. Computer just reboots on its own.

I have absolutely no idea about the causes of these issues. I know for a fact it's not temperatures, as I constantly keep and eye on them and just yesterday replaced the thermal paste on the CPU and the GPU. There aren't any weird capacitors on my motherboard. They all look fine and its a kinda heavy duty motherboard from ASUS, from the TUF line, whose components are supposedly military grade or some marketing crap like that. My power supply is a 650W so maybe it's not that either.

I've browsed a lot trying to find an answer to my problem, and on this website the closest issues to mine had answers like the display returning electricity through the cable to the computer thus making it act weird, which sounds a little preposterous, and something else about hardware. I might be slightly convinced that its the graphics card but not entirely sure. I have updated to the latest Nvidia drivers, if that's of any interest. I have also made sure everything is seated correctly in the motherboard and reset the CMOS with the cap thingy, twice. I also reset the BIOS to default settings, even though they already were that way. Just checked I didn't have any weird voltages or anything. I also ran MBAM and it came up clean.

Here are my specs
-AMD FX 8320 8 core processor, not overclocked and aircooled with a Cryonic H7. Stock clock is 3.7GHz
-ASUS 990FX AM3+ motherboard.
-ASUS GTX660 graphics card, not overclocked.
-Corsair Vengeance Pro RAM 16gb, two sticks of 8gb at 2400Mhz if I'm not mistaken. Not overclocked.
-Corsair TX650W power supply.
-1 Seagate 1TB drive and 1 WD Green 2TB drive.
-Windows 10

Computer is connected to a Samsung Smart TV via HDMI.

I hope you can help me find the solution to this problem. Unfortunately, if it comes to buying new hardware, it's not something I can do so easily. So I guess that if it's a hardware problem, sucks to be me.

UPDATE: I started running games in windowed mode and with a lower resolution. Problem seems to be gone. I'm starting to think it might be a Windows 10 issue, but also keeping my eye on the GPU.

UPDATE 2: Running games in windowed mode didn't fix the problem. I also deleted all Nvidia drivers and downloaded again. Problem still isn't fixed. I checked the event log for the last time it crashed and it shows a Kernel-Power ID 41 problem, which makes me think it might be power/PSU related. I connected the PC directly to the wall, as opposed to the power strip i would normally use. I will also format and reinstall Windows 10 to rule out any software problems. Will keep updating this thread. It's an incredibly frustrating problem and if someone else has it I want to help find a solution.
 

Unfortunately, no. It if is the GPU then I'll just have to deal with it until I can get a new one.
 

Yup. Warranty ship sailed a long time ago. But anyway. Thanks for the second opinion.
 
I haven't entirely ruled out a GPU problem yet, but I'm increasingly confident in that it might be my PSU that isn't working properly. Just yesterday I was checking the BIOS settings when I was that the 12v rail was measuring at 10 volts, and 5 seconds after that, my computer rebooted. I ran 3D Mark and the Tomb Raider benchmark and at no point did the 12v rail actually get to 12 volts. It stayed fairly low. This would explain the Kernel-Power ID 41 issue shown in the event log when the computer loses signal and then restarts.

Here is an imgur link showing the HWMonitor values after putting the GPU at 99% load. Notice the 12v rail values. http://imgur.com/a/KncK3
I don't own a multimeter but I might acquire one, if more of you think this might be a PSU issue, so I can measure the values properly.
 


I finally found the problem. It was the power supply. 8 days ago I connected the multimeter to the 24 pin connector, black probe on black cable for ground, and red probe in yellow cable for 12v rail. Turned on the PSU and the computer and the multimeter read 12.04~ volts. It went down 0.02 or so when gaming, but nothing out of the ordinary.

HWMonitor was also reading higher than usual. Before, it used to hover between 11.6 and 11.3 or so. However, when I have the multimeter plugged in, it reaches 11.952 and higher, never lower, which it didn't do before. I played and used my computer for about 2 days with the multimeter plugged in and ran stress tests for both the CPU and the GPU and I didn't crash even once. I took out the multimeter and the system remained stable for a week. HWMonitror readings were within acceptable parameters, but I noticed they were decreasing by a little bit each day.

Today, my monitor lost signal, and I had to reboot the computer, twice. I plugged the multimeter back in, and guess what, voltages are back to normal. Someway, somehow, plugging the multimeter into the 24 pin makes voltages go back to normal. I thought it was some loose cable that got pushed in by the probe but that wouldn't explain why the readings went down gradually. So it's safe to say that the problem has been the 12v rail on my power supply all along.

I'll be ordering a new power supply today. Hopefully this will go away forever.