Hello,
I recently upgrade from Ryzen 5 3600 to a Ryzen 7 5800x3d and now i'm experiencing random reboots.
It happens when gaming but don't really know what trigger it. I can play for several hours whithout trouble and suddenly pc restart without any error message. It occurs mostly every time i leave game focus (alt+tab or press windows key) or when i launch another program (ex: spotify or chrome) while game is still running in background.
I try to test PSU with CPU+GPU Stress Test (with OCCT) but nothing happened.
Temperature not seems to be the problem, never go higher than 70°C.
The only thing i can find after reboot happen is an error 41 in event log with different bugcheckcode each time, not sure if it can tell something.
Everything is normal as soon as i put the Ryzen 5 3600 back.
My system:
Motherboard: MSI B550 Gaming Edge Wifi (Up to date BIOS and default settings)
CPU: Ryzen 7 5800x3d
CPU Cooler : Noctua Nh-u12s
GPU : XFX Speedster MERC310 7900XT
RAM : 4 x Corsair Vengeance RGB Pro 8Gb 3200MHz
PSU : Corsair RM750x
SSD : Corsair Force MP510
Windows 11 22h2
Does anyone has thoughts on what i can do?
I read that faulty cpu can cause that but is there a way to test?
Thanks by advance for your help
What CPU fan curve are you using? Have you tried setting the CPU fan speed to 100% and try to replicate the issue? Based on the NSPR published by Noctua the Nh-u12s may be operating at its TDP limit. The rest of the system specs seem quite capable but as you can swap the CPU and the issue goes away it definitely would seem related to heat generation or power draw.
The CPU fan is set to 100% when CPU hit 70°C.
I use AMD Adrenalin overlay to watch temp and when reboots happen the CPU is around 65°C.
I tried cpu stress test (cinebench and OCCT) and max temp was 75°C.
If I understood correctly from what I read, if the CPU reaches its maximum of 90°C it should reduces the frequencies, not shutdown instantly like that no?
I also suspect my power supply of being a little weak and not supporting some power peaks. I think I will order a bigger one (850 or 1000W) hoping it's that the issue.
A 750W Corsair should be plenty for your system. 7900XT is a 300W card, the 5800X3D is around 100W. You might see spikes up to 10% above that but you should still have plenty of power available. I'm running a 750W Silverstone with my 7900XT and 5700X overclocked and no problems.
The X3D V-cache is more heat and voltage sensitive. If you have CPU fan set to 100% and you're still hitting 70°C the cooler is likely maxxed out and overrunning the heatpipes. Time for a dual tower with more heatpipes (6 or 7 instead of 5) or an AIO.
I've reinstalled the 5800x3D and check every psu connections and don't get any problems since 3 days now. Didn't notice anything but maybe a connector was not fully plug in.
Next step is to upgrade the cooler to a Nh-u12a as dual tower don't fit with my ram and i'm more confident with air cooling.
Anyways, thanks a lot for your help
Good to hear Pierlorem, possibly reseating the cooler corrected an uneven mounting or thermal paste application. Since you had reported switching back and forth between the 5800X3D and 3600 that would have been done already, but not being familiar with the cooler maybe it can be a difficult installation?
According to AMD Specs on your new Ryzen 7 5800X3D it has a high TDP Rating of 105 Watts and a Maximum Operating Temperature of 90c.
According to Noctua Compatibility Chart for the Ryzen 7 5800X3D your Noctua U112S is strong enough for that processor: https://ncc.noctua.at/cpus/model/AMD-Ryzen-7-5800X3D-1597
According to AMD Specs for the RX7900XT you need a Minimum PSU of 750 Watts to run your GPU card in your PC.
I would Stress test your PSU and GPU and CPU with OCCT and see if it shuts down while being Stress tested.
Keep a close eye on Temperatures, Fan speeds, and PSU Outputs (3.3/5.0/12 VDC).
According to this CPU/GPU - PSU website you will need at least a 650 Watt PSU for your processor and GPU card: https://www.whatpsu.com/psu/cpu/AMD-Ryzen-5800X3D/gpu/AMD-Radeon-RX-7900-XT
But if you have a old or cheap quality PSU it might not be able to output enough power to run your PC. Thus stress testing your hardware with OCCT.
Generally random shutdowns are caused by:
1- Overheating issues
2- Power issues
3- Over/Under clocking issues
4- Incompatible hardware issue
5- defective hardware issues
6- Driver issues
Those are some of the major reasons that comes to my mind at the moment.
NOTE: I would make sure your Noctua CPU Cooler is correctly installed and making a good tight even connection to your CPU surface. Also if it only has one fan add a second fan in a Push-Pull configuration to give extra cooling to the CPU Cooler plus help in having better air circulation inside your PC.
If you have poor air circulation inside your PC it can cause your hardware to overheat. You can check by removing your PC Side panel and if the CPU temperature is slightly less than with the panel on that indicates poor air circulation in your PC.