[SOLVED] Brand new PC freezes under load

Keny2323

Commendable
Feb 14, 2017
11
2
1,515
I've assembled a new PC like 3 weeks ago, I had a stable OC running on my vega 56 (never had any crashes) until I installed 3 case fans. Ever since I installed them my PC is completely unstable. It can go as much as 1 day without freezing, sometimes it can't even last an hour. After unplugging the case fans the issue still remains.
Things I tried:
-Reseating every component, including cables
-Switching DIMM slots
-DDU-d drivers and installed older ones
-Fresh install of windows
-chkdsk showed no errors
-memtest86 showed no errors
-I'm using 2 separate 8 pin PCIe cables instead of 1 daisy chained one
-Plugging the psu cable straight into the wall instead of an extender
-Updating windows
-Updating chipset drivers
-Temperatures are around 60 celsius for the cpu under load and 75 for the gpu.
-Removing every OC (even the default boost clocks)
-Undervolting my gpu
-Stress tested my GPU for 20 minutes without any crashing
-Stress tested my CPU for 15 minutes without any crashing
-Eventvwr recognises an improper shutdown but gives no useful information

Despite all the effort my computer still freezes under load. Sometimes I can hear a buzzing sound, sometimes it's a completely silent freeze. Also, my most recent freeze was quite interesting. My game froze but I could still move my mouse for about 1-2 seconds after which the mouse froze too and the system became unresponsive. When it freezes, restarting the PC will not allow it to boot, it will only be a black screen even though every fan is running. In order to boot it up again I have to shut it down by holding the power button and then starting it again.

My specs are the following:
MSI Vega 56
MSI b450m pro-m2
Corsair Vengeance DDR4 16gb 3000mhz
Ryzen 5 2600
Corsair TX650M 650W 80+ gold PSU

Please someone help me with this, it's making me go nuts. Thank you in advance.
 
Last edited:
Solution
Okay, I've got a solid lead now. I found a post on the amd forum about the same issue that I'm having and guess what, the guy has the exact same psu that I have. Do you think I should return the psu and get a new one? EDIT: the guy replied and said that it was his card that was faulty. I'm returning my card to amazon.
Yes, it will go back to my previous point that it is likely the GPU. So returning it will find out. Let us know if it works :)

Keny2323

Commendable
Feb 14, 2017
11
2
1,515
Do you know where the buzzing sound is coming from?
Does the freezing still occur in safe mode?
Have you reset CMOS just to ensure the BIOS is completely wiped to default?
Is the PSU new?
The buzzing sound comes out of my headphones.
I have not tried safe mode, will give it a shot. Thanks
Yes I have reset CMOS, my bios is on default settings.
 

PC Tailor

Illustrious
Ambassador
I can't test safe mode. There is no way I can stress test the system as most games run terribly in safe mode.
Games aren't meant to be played in safe mode, as safe mode disables all other drivers.
So the freezing only happens under heavier loads / whilst gaming?
How many passes did you run memtest for?
Do you have latest BIOS and all latest MB/Chipset drivers installed?
Have you tried another GPU to verify that the OC hasn't caused a problem?
 

Keny2323

Commendable
Feb 14, 2017
11
2
1,515
Games aren't meant to be played in safe mode, as safe mode disables all other drivers.
So the freezing only happens under heavier loads / whilst gaming?
How many passes did you run memtest for?
Do you have latest BIOS and all latest MB/Chipset drivers installed?
Have you tried another GPU to verify that the OC hasn't caused a problem?
Yes, the freeze only happens during load.
I've ran 4 passes.
The latest BIOS is actually a beta bios and it's for new ryzen cpus so I doubt it will make a difference.
Aside from that, I do have the latest chipset/mb drivers.
I can't try another GPU sadly as my old system was a laptop. (btw I tried power limiting to -30% so my peak wattage was 110W and I still crashed :/)
 

PC Tailor

Illustrious
Ambassador
I can't try another GPU sadly as my old system was a laptop. (btw I tried power limiting to -30% so my peak wattage was 110W and I still crashed :/)
That would be potentially useful if the PSU was faulty, but less indicative if the GPU is faulty.
If you can, it may be worth trying a different GPU and seeing if the issue persists.

PSUNew - not tested????
GPUNot tested - could use Furmark????
RAMRan memtestShowed no errors (4 pass)
MotherboardNot tested????
CPUNot tested - Could run prime95????
StorageNot tested????
OSClean installIssue persisted.
BIOSUpdated BIOSIssue persisted.
SoftwareUpdated all driversIssue persisted.
PeripheralsNot tested????

It may be worth checking the status of your storage drives. I've covered general aspects above. There's no guarantees. Memtest can still pass and the RAM can still be faulty for example.

The fact that you have updated BIOS and have clean installed Windows again would indicate it's more likely hardware. Being as it came out of the back of a GPU OC, I'd be tempted to try and borrow another GPU after the storage drive check.
 

Keny2323

Commendable
Feb 14, 2017
11
2
1,515
That would be potentially useful if the PSU was faulty, but less indicative if the GPU is faulty.
If you can, it may be worth trying a different GPU and seeing if the issue persists.

PSUNew - not tested????
GPUNot tested - could use Furmark????
RAMRan memtestShowed no errors (4 pass)
MotherboardNot tested????
CPUNot tested - Could run prime95????
StorageNot tested????
OSClean installIssue persisted.
BIOSUpdated BIOSIssue persisted.
SoftwareUpdated all driversIssue persisted.
PeripheralsNot tested????

It may be worth checking the status of your storage drives. I've covered general aspects above. There's no guarantees. Memtest can still pass and the RAM can still be faulty for example.

The fact that you have updated BIOS and have clean installed Windows again would indicate it's more likely hardware. Being as it came out of the back of a GPU OC, I'd be tempted to try and borrow another GPU after the storage drive check.
Actually I did test my GPU with furmark, ran it for 20 minutes without any crashing.
I also ran prime95 for about 15 minutes, again, no crash
 

PC Tailor

Illustrious
Ambassador
Actually I did test my GPU with furmark, ran it for 20 minutes without any crashing.
I also ran prime95 for about 15 minutes, again, no crash
Technically prime95 should be run 12/24 hours. I see CPUs pass 60 minutes and fail hours later.
Now you don't always need to run it this long, but it should be more than 15 minutes definitely.
Again, they're just ideas, Furmark definitely isn't a guaranteed, but it's just an identifier for any obvious instability.

PSUNew - not tested????
GPUFurmark for 20 minsNo crash
RAMRan memtestShowed no errors (4 pass)
MotherboardNot tested????
CPUNot tested - Could run prime95????
StorageNot tested????
OSClean installIssue persisted.
BIOSUpdated BIOSIssue persisted.
SoftwareUpdated all driversIssue persisted.
PeripheralsNot tested????
 

Keny2323

Commendable
Feb 14, 2017
11
2
1,515
Technically prime95 should be run 12/24 hours. I see CPUs pass 60 minutes and fail hours later.
Now you don't always need to run it this long, but it should be more than 15 minutes definitely.
Again, they're just ideas, Furmark definitely isn't a guaranteed, but it's just an identifier for any obvious instability.

PSUNew - not tested????
GPUFurmark for 20 minsNo crash
RAMRan memtestShowed no errors (4 pass)
MotherboardNot tested????
CPUNot tested - Could run prime95????
StorageNot tested????
OSClean installIssue persisted.
BIOSUpdated BIOSIssue persisted.
SoftwareUpdated all driversIssue persisted.
PeripheralsNot tested????
Okay, I've got a solid lead now. I found a post on the amd forum about the same issue that I'm having and guess what, the guy has the exact same psu that I have. Do you think I should return the psu and get a new one? EDIT: the guy replied and said that it was his card that was faulty. I'm returning my card to amazon.
 
Last edited:

PC Tailor

Illustrious
Ambassador
Okay, I've got a solid lead now. I found a post on the amd forum about the same issue that I'm having and guess what, the guy has the exact same psu that I have. Do you think I should return the psu and get a new one? EDIT: the guy replied and said that it was his card that was faulty. I'm returning my card to amazon.
Yes, it will go back to my previous point that it is likely the GPU. So returning it will find out. Let us know if it works :)
 
Solution

PC Tailor

Illustrious
Ambassador
Lower the engine clock of your GPU card to see if it is the PSU that can not deliver enough power.
This will still not guarantee if the PSU is faulty, as really it only needs a rail to fluctuate +5% voltage to cause an issue, or some ripple to cause an issue either at load or not. You can always test this with a multimeter without load, but it's not guaranteed, so returning the GPU is the best bet being as it has followed a GPU overclock, then we can see.