Random restarts, sometimes CPU debug led on

[Qwetaa]

Hello everyone!

I have been enjoying my build for a year now and of course, problems arised. Of course, it's in happening when I don't have money and time for servicing/warranty...

Rig:
I5 6500 (stock cooler)
MSI Z170A PC MATE
RAM 2x8 GB HyperX 2133 MHz
3 HDD, 1 SSD
Radeon RX480 HD
Tagan 600 W (few years old PSU)

It all started when I was playing Fallout 4 month ago - random restarts after a few minutes of gameplay. I decided to clean my PSU - problem solved for 2 weeks. After that - restart, no POST, EZ Debug LED on CPU. I replaced thermal grease and noticed slightly (no cracks) bent CPU bottom corners, on the other hand socket is intact. I turned on PC - EZ Debug LED CPU was on again, but after 3 minutes, computer booted.

Now crashes are really random - sometimes after a day, hour and sometimes with debug led...
Temperature readings are okay (70 on GPU and HDDs, 50 Celsius on the rest of components), I also measured voltages on PSU under load (benchmark) - 11.9 V and 5.3 V, I wasn't checking 3.3 V rail.

Windows log shows Kernel Power Failure.

Any ideas?

 

[Kenton82]

Is that 70 degrees CELSIUS on your GPU - AND HDD?? They are too high for a start, and that PSU needs swapping out ASAP! A couple of potential problems there. Bad PSU and possible thermal throttling. If the PSU is starting to go bad, it can also trigger shutdowns.

 

[Qwetaa]

As for temperatures, I feel a need to clarify that these are NOT idle temps and from what I read a couple months ago, it's normal for RX480 to stay on 70C so I would discard that problem.

As for PSU - I feel like that might be it (especially concerning that 2 weeks without problems after cleaning PSU radiator), but that CPU led is troubling me - CPU is slightly bent, but works, could it be that sometimes it loses connection? How? And why now? Or could it only be CPU going kinky because of indeed faulty PSU?

Anyway, I don't have spare PSU to check it, but I think I'll buy new PSU anyway, I just don't know should I pick more powered one - like 650/700 W or just stick to 600 W.
If it's not PSU then I think I'll have to pray for CPU warranty and if it's void, play with heat gun :/ What do you think?

 

[Kenton82]

Running the card at 70oC could be improved, as it is on the warm side period. The bent pin COULD be a problem, as the CPU heats up and cools down, it will e expand and contract ever so slightly, so this MAY be a problem. Hard to tell unless we can see it in action. THe PSU does need swapping out ideally, and no, you do not need more power, 600w would be ample. But do get something of quality. Corsair RM***X series, or EVGA G2 (or newer) or if you can afford it, something from the Seasonic range.

The Kernel power failure would point to the PSU

 

[Qwetaa]

Thanks for the replies Kenton. As for PSU, I agree that it's not the thing to "save money on". I will just buy new PSU on monday and we will see if that helps.

 

[Kenton82]

Thanks for the replies Kenton. As for PSU, I agree that it's not the thing to "save money on". I will just buy new PSU on monday and we will see if that helps.

Ok good start, do update the thread with your findings.

 

[Qwetaa]

Hey, a little update from me - last crash on day 05.08.2017.
I am still waiting for my salary to buy and replace PSU. As for now, I tried running memtest (2 passes without errors with both sticks inside), torture tests from Prime95 and additionally Unigine Heaven Benchmark 4.0 simultaneously with Prime95 torture tests + web browsing. I wasn't able to even get my pc to overheat... Well, restarts occured even when PC was idle but I guessed it was worth giving a try to completely exclude temperatures.
There is no visible damage anywhere apart from cpu which I mentioned earlier, but I wouldn't be surprised if after running Prime95 CPU straightened itself under stock cooler pressure xD (temp was 80-85 max because of sh***y cpu fan controlled probably with internal thermistor).
Now I am completely confused - if everything is seemingly okay, why there were sometimes problem with POST after reboots with CPU debug (but fans were spinning).
Problem started when PSU and it's air filter were clogged with dirt to disappear for 2 weeks after cleaning. What irritates me most is that I can't reproduce faults so I could point precisely to the root of my problem.

 

[Qwetaa]

Swapped PSU with brand new EVGA 600 W. Before swap there was no crash in 3 days. In the meantime I have checked RAM and its sockets, installed new OS. It would take a month for me to confirm solution, as crashes are random, but now with new PSU, OS and 2 fans on top of case I have lowered temperature on HDDs to max 58 C.
Elder PSU was 5-7 years old and it was exposed to a few short-circuits in USB Port on previous MOBO (messing with breadboard circuits) and voltage spikes from a motor of old dormitory fridge, thus it would fail anyway soon IMHO.
Problem seems to be solved as for now.

 

[Kenton82]

Swapped PSU with brand new EVGA 600 W. Before swap there was no crash in 3 days. In the meantime I have checked RAM and its sockets, installed new OS. It would take a month for me to confirm solution, as crashes are random, but now with new PSU, OS and 2 fans on top of case I have lowered temperature on HDDs to max 58 C.
Elder PSU was 5-7 years old and it was exposed to a few short-circuits in USB Port on previous MOBO (messing with breadboard circuits) and voltage spikes from a motor of old dormitory fridge, thus it would fail anyway soon IMHO.
Problem seems to be solved as for now.

Thats great news, just keep an eye on it. The temp of the HDD still seems very high though....
Keep an open mind of a new setup!
K.