PC randomly crashes

[Incognito Clown]

Last month, my PC has crashed 3 times: it instantly shuts down (not a windows shutdown, everything just goes black, with the exception of the motherboard rgb's). This happened during gaming (COD and F1). However, I pretty much game every day between 3-4 hours a day, and it "only" happened 3 times, so I cant really tell if it only happens while gaming. It kind of looks like a sudden loss of power, but my monitors and router stay on and are connected to the same power outlet, so it is not a power outage.

 

Entries of "Kernel-Power" in Windows Event Viewer:

• 17/05/2019, 19:56:20 - The system has rebooted without cleanly shutting down first. This error could be caused if the system stopped responding, crashed, or lost power unexpectedly.
• 29/05/2019, 19:30:27 - The system has rebooted without cleanly shutting down first. This error could be caused if the system stopped responding, crashed, or lost power unexpectedly.
• 10/06/2019, 13:18:34 - The system has rebooted without cleanly shutting down first. This error could be caused if the system stopped responding, crashed, or lost power unexpectedly.

I've never had such an issue before may 17th.

 

(possibly) related info:

• I've upgraded my motherboard and CPU on march 8th (about 40 days before the problem occurred for the first time). During this upgrade I also rebuild/reseated everything else in my PC.
• My CPU is undervolted, but does pass stress tests and does not throw WHEA errors (hwinfo always runs in the background and would notify me of these).
• My power supply is a Corsair RM750i which is about 2 years old. It should be able to handle up to 750W, and I've never seen my peak power draw anywhere above 430/450W, so it should be sufficient.
• I've never seen CPU temps above 73 degrees Celcius while gaming. However, I've not been able to check the temps at the exact moment of the crash, since I can obviously not see it coming.
• GPU temps are normally sitting between 68 and 72 degrees Celcius. However, I've hit 79 degrees in the past on a hot summer day, when playing a racing game for multiple hours on multiple screens. This never caused any issues, and the previously mentioned crashes where not on such hot days.
• Running stress tests like AIDA64 for an hour do run normally. CPU temps then do go up to around 85 degrees, but this still does not cause the system to crash or even throw whea errors.
• The crash seems to happen instantly. The game does not slow down or freeze up and fans do not start or stop spinning just before the crash.

System info:

• i9-9900k at default clocks, but undervolted (1.280V in bios)
• Gigabyte Aorus Ultra Z390 mobo
• Hydro h100i 2x120mm CPU cooler, as intake in the front
• 3x noctua case fans, as exhaust
• 1080TI, no custom overclock
• 2x8GB ram, XMP 3000 profile
• 2x sata ssd, 1x sata hdd, 0x m.2

 

Does anyone have any suggestions to what the cause of the problem could be, or what my next step should be?

[paddy-stone]

Well, my first step would be to reset CMOS and run default settings for a while to see if the problem occurs. If the problem doesn't occur, then I'd hazard a guess that it's either the RAM settings or undervolting that caused the failure. Although saying that, it sounds like a power fault to me, which is why it reboots/turns off rather than showing a BSOD.

[Valkyrie Lenneth]

search for bccode in event viewer, at the crashes , that will give alot of info

[Incognito Clown]

Thanks for your suggestion @paddy-stone. I've clicked "Load optimized defaults" in the Bios. My undervolting is now reset and the memory is running at JDEC speeds instead of XMP. Does this accomplish the same as a CMOS reset by taking out the battery (which is conveniently placed behind the gpu...)?

 

@Valkyrie Lenneth I've not gotten a BSOD. Therefore I don't see a bccode in the event viewer.

[paddy-stone]

40 minutes ago, Incognito Clown said:

Thanks for your suggestion @paddy-stone. I've clicked "Load optimized defaults" in the Bios. My undervolting is now reset and the memory is running at JDEC speeds instead of XMP. Does this accomplish the same as a CMOS reset by taking out the battery (which is conveniently placed behind the gpu...)?

 

@Valkyrie Lenneth I've not gotten a BSOD. Therefore I don't see a bccode in the event viewer.

Yes, essentially the same thing. Should be good for some troubleshooting now, to see if it still borks.

[Incognito Clown]

@paddy-stone Alright I will run the PC like this for a while. Will report back here in a month or when the problem occurs again

[Valkyrie Lenneth]

21 hours ago, Incognito Clown said:

Thanks for your suggestion @paddy-stone. I've clicked "Load optimized defaults" in the Bios. My undervolting is now reset and the memory is running at JDEC speeds instead of XMP. Does this accomplish the same as a CMOS reset by taking out the battery (which is conveniently placed behind the gpu...)?

 

@Valkyrie Lenneth I've not gotten a BSOD. Therefore I don't see a bccode in the event viewer.

most likely psu or ram or mem controller in that case

[Incognito Clown]

@Valkyrie Lenneth @paddy-stone

 

After 21 days without any crashes, I've got one today... This was without the CPU under volt applied, with the JDEC mem spec (not xmp) and also with the GPU temp+power slider reset to auto (was 116% / 90 degrees).

 

  List of the crashes so far. "Sata ext" stands for a sata power extension cable I'm running to my sata SSD's (2). Thats besides the mobo/CPU the only new part.

 

My pc runs fine for weeks. Even when gaming on it for 4-5 hours a day. The motherboard and the CPU are the only "new" things in the pc, but if they would be broken, I'd expect way more issues. Not once every 2-3 weeks.

I'm not really sure how to proceed. Any ideas?

One thing that came up is that maybe it could be my Corsair AIO. I know that it gets power via the 4 pin fan header. What if it for some reason does not get enough power for a few seconds which would then cause its pump to stop spinning and the CPU temps to climb rapidly. I always have msi afterburner open, but I expect temps to climb too fast in such case to actually be able to notice it before it thermal-shuts down. However, I did not notice any lag spikes before the crash (caused by thermal throttling), but then again, I've no idea how fast the CPU overheats if you give it 0 cooling for a few seconds.

[paddy-stone]

Hmm, that is a strange one. Not sure TBH where to go from here in your case. I will say however that how you've described the fault sounds almost exactly like when I had a PSU dying on me... I also got random reboots of the system for a looong time and was a bitch to diagnose it. In the end it completely failed and went with a bang while I was using it in my test system.

 

If it were me, I'd swap parts out one at a time to see if it still does it... but I know not everyone can do this.

[Incognito Clown]

@paddy-stone

 

Well, I do have another desktop PC which I use as my NAS. I could swap its Corsair CX650M with the Corsair RM750i thats currently in my game PC and see how it performs.

 

If that does not work I can also swap the memory sticks. As a last resort I can also take my old (pre-upgrade) motherboard and CPU and run that for a while, but I'd rather skip that one  I sadly do not have a second CPU cooler which would be sufficient for the i9. Some of those air coolers are not too expensive though, so I'd not mind to buy one just to try it.

[Incognito Clown]

@paddy-stone

 

Alright, I just found out the pump of my H110i v2 is reporting that its spinning between 1,290 RPM and 1,560 RPM. In the iCUE profiler it is set to extreme (which should be 3,000 RPM). The 2 other profiles (quiet and balanced) should let the pump spin at 2,000 RPM. I've not seen such low speeds before, and I've checked them every day last 3 weeks. It also no longer reports the fan speed (although they are spinning).

 

The problem persists even after rebooting the system.

[Incognito Clown]

@paddy-stone

 

Turned out the fan header was not set to a fixed value of 100% (anymore?). Fixed that, issue with the AIO solved.

 

I also came up with the idea to make hwinfo log its values to a log file, which already paid of (sadly) because my pc crashed twice this evening (2nd time not logged)... CPU temps were perfectly normal the package hit 90 degrees for a second during the 4 hour game session and never went any higher. Out of all the logs, I've only seen one thing that scared me a little; the psu temps and its fan speed.

 

The last log entry (so up to 1-2 seconds before the crash) logged the following:

PSU Temperature [C]: 69

PSU Temperature2 [C]: 47.5

PSU Fan [RPM]: 0

PSU Power (sum) [W]: 363.54

PSU Efficiency [%]: 91.6

 

Those temps where both the highest temps reported during the 4 hour game session.

 

Earlier, the fan of the PSU did start spinning at 50.5 (temp sensor 1) and 37.3 (temp sensor 2) degrees.

 

After the crash I booted the system again and it stopped after 2 seconds again.

After 30 seconds I retried, it did start, but after 6 minutes in-game it crashed again.

 

This all makes me think the PSU overheated...

 

 

[Incognito Clown]

Thanks @paddy-stone and @Valkyrie Lenneth. It turned out to be my PSU after all.

I switched the power supply in my gaming rig (the faulty one) with the one from my NAS. Both systems ran fine after that for 1,5 months. When swapping them back my gaming pc crashed again within 2 days. At that point I was pretty sure the issue was with the PSU. The NAS is on 24/7 but uses way less power than my pc (<50w I'd guess). Seems that the PSU was just no longer able to reliably handle higher loads.

 

I sent the PSU in for warranty to Corsair. They sent me a new one within a week. Great service 

The new one has been working flawlessly for a month now 

[paddy-stone]

15 hours ago, Incognito Clown said:

Thanks @paddy-stone and @Valkyrie Lenneth. It turned out to be my PSU after all.

I switched the power supply in my gaming rig (the faulty one) with the one from my NAS. Both systems ran fine after that for 1,5 months. When swapping them back my gaming pc crashed again within 2 days. At that point I was pretty sure the issue was with the PSU. The NAS is on 24/7 but uses way less power than my pc (<50w I'd guess). Seems that the PSU was just no longer able to reliably handle higher loads.

 

I sent the PSU in for warranty to Corsair. They sent me a new one within a week. Great service 

The new one has been working flawlessly for a month now 

Glad you got to the root of it. It was always my first choice of what the problem was, but as you know with PCs it can be hard to diagnose sometimes.