Question New PC randomly deadlock and needs to be force restarted.

[HelpMeI'mHopeless]

I built a new PC and almost every day it freezes randomly. Usually when it freezes I only have Chrome open. Doesn't happen during gaming sessions. Everything becomes unresponsive. Caps lock button doesn't light up. I need to hold down the power button and restart whole PC. There is no pattern, it just happens every couple of hours.

I ran diagnostics for ram, ssd, gpu, everything was fine.
I updated motherboard bios and that seemed to fix it but the issue came back.
I tried with xmp enabled as well as disabled, no difference.
Event viewer doesn't say anything. Windows doesn't create mindump files.
All drivers are up to date as far as I can tell. (If you know a better way of checking, please tell)
My power plan is set to most performance.

This has been happening pretty much since the day I built the pc. What else could I try before completely reinstalling the windows/drivers? I finally finished setting up environment for programming work, so I would rather not have to delete everything and start from scratch.

My pc specs:
CPU - intel i5-12400f
GPU - AMD RX 6600
MBO - Gigabyte b660m gaming x ax ddr4
RAM - Teamgroup tforce vulcan z
PSU - EVGA g5 650 watt

Any help will be greatly appreciated.

 

[HelpMeI'mHopeless]

Your symptoms seem consistent with a cpu stop.

What's your cooler? Is it properly attached with thermal paste? Monitor temps with hwmonitor

Please also read vCore value or post a screenie of hwmonitor so the power outputs are also visible.

Could be bent contact in the mobo socket: inspect with magnifying glass for any misaligned contacts.

Thanks for trying to help.
Right now I am running Prime95 test. It's been up for good 15 minutes and max temp is 67, average is 56. My cooler is ID-Cooling SE-224XT. I think I installed it and applied thermal paste properly.
From HWiNFO64 - Vcore values:
Current 0.324V
Minimum 0.072V
Maximum 1.116V
Average 1.002V
When I was building, pins were fine. But if i have no other choice, I will take it apart and check again.
If you need any more info, pls ask. Maybe mods will delete this thread, so feel free to continue in private conversations! Thanks again.

 

[HelpMeI'mHopeless]

Thanks for trying to help.
Right now I am running Prime95 test. It's been up for good 15 minutes and max temp is 67, average is 56. My cooler is ID-Cooling SE-224XT. I think I installed it and applied thermal paste properly.
From HWiNFO64 - Vcore values:
Current 0.324V
Minimum 0.072V
Maximum 1.116V
Average 1.002V
When I was building, pins were fine. But if i have no other choice, I will take it apart and check again.
If you need any more info, pls ask. Maybe mods will delete this thread, so feel free to continue in private conversations! Thanks again.

@DavidM012

 

[DavidM012]

A stress test according to the overclocking section of the forum should be able to run for hours so, how can it complete a stress test if it's stopping every couple of hours?

Are all your power cables correctly attached and the ones that shipped with the power supply? Otherwise, everything looks ok.

If it isn't the socket I can't think of anything else immediately.

 

[DavidM012]

Maybe the cooler is unevenly tightened, or over tight. That can cause 'ghost' problems.

 

[HelpMeI'mHopeless]

A stress test according to the overclocking section of the forum should be able to run for hours so, how can it complete a stress test if it's stopping every couple of hours?

Are all your power cables correctly attached and the ones that shipped with the power supply? Otherwise, everything looks ok.

If it isn't the socket I can't think of anything else immediately.

I haven't tried running it for hours. Should I do it?
Cables are correctly attached as far as I know. But I might try taking it apart and plugging everything again.

 

[DavidM012]

You could run the stress test as long as you like, if the system is well cooled it should hold up if it's healthy. However since the system we know is stopping every couple of hours there doesn't seem much point as the result seems predictable at this stage.

 

[HelpMeI'mHopeless]

Maybe the cooler is unevenly tightened, or over tight. That can cause 'ghost' problems.

Pretty sure I tightened it evenly. Overtightened idk. But something is telling me that it's a software related issue.

 

[DavidM012]

Windows 10 or 11?

 

[HelpMeI'mHopeless]

win 10, did all the updates

 

[DavidM012]

Well the only other thing I can think of, I've seen a similar case recently where everything looked normal and diagnostics said the disk was ok, but switching the disk solved it.

However what you could do is turn off your screensaver and run the stress test for a few hours, don't move the mouse or open any apps and see if it's still running in say 3 hours or any period of time longer than the time you've seen already.

Keep an eye on the temps. to ensure it's within safe margins. You could also open task manager and see if the cpu usage dips under 100% during the test while it's going.

 

[HelpMeI'mHopeless]

Sadly I don't have another drive to switch the ssd.
And btw, just had another freeze moments ago.
I will update to windows 11. If this issue persists I will reinstall everything from scratch.

 

[DavidM012]

It's basically a process of elimination. No harm installing windows 11. If the problem still persists with windows 11, then it's likely some hardware problem. The OS runs on the disk. Any problem would disrupt the system.

If you run the stress test it'll tell you something if it's still running whilst unattended.

There are more than 1 suspects in the case: The cpu & cooler being properly seated, the drive or drive port, you can try switching the disk to the other m.2 port, and checking to see it's properly seated.

You think software, so what did you install? Sounds like a complex stack of something.

For example see how the system runs with only windows installed and nothing else, immediately after setup and installing all the drivers. Could try installing the mobo chipset drivers, if you haven't done so already, you said you installed all drivers up to date though.

Also some ssd manufacturers have an n.v.m.e driver specific to their product, see if they have one of those.

 

[PEnns]

Usually when it freezes I only have Chrome open.

Does it happen even when Chrome is not running?? And if yes, how many tabs are open??

 

[HelpMeI'mHopeless]

Does it happen even when Chrome is not running?? And if yes, how many tabs are open??

Yes, even without chrome

 

[HelpMeI'mHopeless]

Here's the update.
I upgraded to win11, still deadlocking. So i went and got win10 installation media and clean installed win10, formatted ssd too. Before that, i opened the case, reseated the ram sticks, made sure everything was plugged in and cooler screws were balanced. However, bad news...
Haven't even installed the drivers and already having the same issue, so pretty much guaranteed hardware related. I had nothing open, just moving my mouse around every two mins and pc just froze. Nothing responds not keyboard not mouse. I have no idea what it could be. I don't have any spare parts to test anything. No drive, no old gpu, not old ram. I'm at a loss about how to proceed. Hopefully it's something minor. I will install drivers/update bios and see how it goes. But I just don't see it happening. Any help would be appreciated!

 

[DavidM012]

It's rare but maybe your cpu is faulty out of the box. The cpu does seem to be stopping - if the numlock led on keyboard doesn't respond, and the mouse pointer freezes, it's definitely a symptom of cpu stop. Your temps and voltages are ok so maybe it's duff in another unusual way.

Last ditch before you start to rma parts is to try the system with one or other of the memory modules, but not both. Can also try memtest86 memory benchmark test to be thorough: if even that crashes it has to be the cpu.

 

[HelpMeI'mHopeless]

one more thing, in m.2 slot there appears to be an led and it is flashing red on the motherboard. Could this mean that something is wrong? It has been like this since I built it and I couldn't find any information in the manual.

 

[DavidM012]

Undocumented led blinking on mobo - have to ask gigabyte. I would consult the manual.

The cpu should keep on going even if the drive isn't working properly. In windows it would crash since windows accesses the drive for virtual memory and other stored data.

If you run memtest86 what's in use is the usb ports and drive (and kyb+mouse i/o) which isn't connected to the pci-e lanes so, the system would be isolated to using usb, cpu, memory and video output and if it can keep running through that, if the cpu was bad even the memtest86 app would crash.

There are more than 1 suspects in the case so, there are indications that the cpu is suspect and indications that the drive is suspect.

That could also mean mobo is suspect. So tests may help eliminate suspects and isolate the cause so you know which component(s) are faulty.

If you remove the drive altogether and run memtest86 this isolates it even more. Does it still blink with the drive not installed? Does memtest86 crash?

If memtest crashes with the barebones system it has to be the cpu. If it keeps on going and completes the test, mobo and drive are suspect.

 

[DavidM012]

You could also run memtest with only 1 dimm installed so it completes faster since it's only for diagnostic purposes.

 

[HelpMeI'mHopeless]

Thanks for comprehensive answer.
I don't understand how running memtest would narrow down our suspects. Can the test run without the drive? Or without gpu? As i dont have another drive available right now (could borrow from friend in the future). Let's say I run the test and it freezes as usual. What next? The fault could be on cpu, ram, mobo, gpu (because cpu is f series, so can't run without it), psu. Because of this I don't understand how running this test would help me.
I agree about trying to run the system only one one stick of ram. I will try this when I get home.

 

[DavidM012]

You have a usb drive. Memtests boots off usb drive. It runs off of it. You can leave the gpu in. It does not run without gpu.

Take the system drive out. That isolates the drive at least.

So... all the system is using is barebones usb, cpu, mem, basic video, kyb and mouse. If this very basic program Crashes then it has to be the cpu.

It doesn't have all the sophisticated layers of windows or run off the nvme port drive. So the cpu should not have a problem running it.

Testing the memory will stress the system a bit but that's no different from prime 95. We want a bit of stress to know it can hold up.

So if it does not crash while the drive is not present, the problem is either the drive or the board. This system is not accessed during the test.

Grant you that the problem could be CPU, drive, board, or a combination thereof:

If you'd rather take shortcuts than run tests you need another drive, mobo, and cpu to simply substitute them one at a time. Don't have. So either RMA parts, ask gigabyte, and test.

I haven't officially suspected the memory yet... memtest will detect errors. If it finds something it'll tell you.

 

[HelpMeI'mHopeless]

Ok, now I understand that it will at least rule out the drive, so that's good. I will research how to run the test and will update you. Thanks again!

 

[DavidM012]

You make a bootable usb similar to windows setup. Insert drive and boot. Just run the memory benchmark and see if it completes or crashes. It will take less time with only 1 dimm installed.

 

[HelpMeI'mHopeless]

You make a bootable usb similar to windows setup. Insert drive and boot. Just run the memory benchmark and see if it completes or crashes. It will take less time with only 1 dimm installed.

Is it necessary to take out ssd? If it boots of the usd, maybe ssd can be left in? In case of failed test, could we still rule out the ssd if it's left in?