PC Processors

[threeheadedsheep]

AMD,

I upgraded my setup from Intel based to the AMD Ryzen 7 3700x.  I've had it for about six months now and after the same issue, I'm here on the forums seeking help. Typically I can diagnose issues like this as I've been working with builds for 15+ years but this one has me stumped. 

Long story short, I'm on my second power supply replacement and I'm about to get a third due to this issue. 

When I initially got the setup running in January parts I reused were the power supply and SSDs. Everything else was new, no known issues with the previous hardware. I gutted the Intel and threw everything together got the PC running, removed old drivers and installed new ones and checked the registry. System is water-cooled on both the CPU and GPU side, no overclocking. Everything was running smooth for about a month and then I started getting random crashes, bsod, no dump file, nothing. It would just shut off briefly and restart. (some history, this is my luck whenever I try to do an amd build... I don't know why but this happened to me the last time I tried an all amd build 10 years ago). 

So the restarts started happening, couldn't figure out why so I'd just work with it. A week passed and then it got to a point where it appeared the PSU completely failed and would no longer restart. Diagnosed every part by swapping and bench testing before I came to the conclusion that it was the power supply. 

I replaced the power supply and it's now end of July and everything was running smoothly again up until last week. The restarted started again and yesterday it finally died on me again. Exact same situation. This time around however, just to be sure it wasnt anything on the software side I reinstalled windows and did a fresh install. It reset 25 minutes into the desktop with nothing running but windows 10 in it's initial state. No bsod, no dump file, system event viewer just shows system power shutoff incorrectly...

So my question is not what failed, as I've already diagnosed it, but why. I don't want to keep buying new power supplies. Should I start looking at the Mobo as a culprit for premature PSU failure? 

System:

AMD Ryzen 7 3700x CPU

ASROCK X570 Mobo

Corsair vengeance 4x8GB RAM

2x SSD

Radeon 5700XT GPU

1st PSU - 1000w Corsair 

2nd PSU - 1250w G.skill 

Timeframe is January - July 

[elstaci]

Users here at AMD have been mentioning about the RX5700 crashing the computer.

You might want to check on that. By either installing the latest AMD Driver or Previous AMD Drive or by installing a different GPU card and see if it continues to crash.

Download OCCT which Stess tests the CPU, GPU, and PSU. See if your computer crashes during the various Stress Tests.

The PSU Stess test is a good test to check out your PSU under very heavy loads.

As for the PSU Failures:

Have you checked your power from the wall outlet?

Maybe it is unstable causing your PSU to fail due to fluctuating voltages. 

One way to make sure your Wall Outlet power won't harm your PSU is by installing a UPS (Uninterrupted Power Supply) which will send a clean stable voltage to your computer.

Also make sure your Wall Outlet is grounded.

Otherwise start looking for intermittent Shorts/Grounds on the Motherboard that can be shorting out your PSU. Like a loose screw or part of the computer case shorting out or grounding the motherboard or some other hardware.

If you did a clean Windows Installation without installing any 3rd party software and it crashed or shutdown then that is a good indication it is hardware related.

[threeheadedsheep]

I ended up taking the computer to an independent PC shop to further diagnose it. They said all the components passed bench tests. They suggested redoing the water-cooling loop on the computer as the pump "was causing the random restarts" when temperature peaked. I don't know how accurate that is but I purchased a new pump and am waiting to redo the loop. 

If the same thing happens after I replace the loop I will be getting a new motherboard. They have not been able to identify any issues with the hardware or software at this time. 

[elstaci]

Overheating CPU will definitely shut down a computer.

After replacing the loop and the CPU & GPU temps are within normal under heavy loads.

Hopefully that will fix your crashes. But if it doesn't then I suggest you open two Support Tickets. One with ASRock and one with AMD Warranty.

A defective CPU will also cause the computer to crash besides a defective Motherboard. You can open a Warranty Request (https://www.amd.com/en/support/kb/warranty-information/rma-form ) with AMD just to see if they believe your problems might be due to the Processor. Same with ASRock.

No harm in asking before spending a lot money.

[threeheadedsheep]

Will a temperature derived reset throw a BSOD? Or is it normal for it to just reset the system without warning?

[elstaci]

Yes, Overheating hardware, Hardware failure or failing, Incompatible or bad drivers, Overclocking and many other reasons will cause BSOD to occur instead of Shutting down the computer.

When a CPU Overheats past the point where it might get damaged it will shut the computer down immediately BEFORE damage occurs.

Normally, when a computer shuts down by itself, like pressing the Power or Restart buttons, is due to these 12 reasons from this Tech site: PC turns off without warning - what to do? | Fixed - Read Our Articles and Optimize Your PC for Peak... 

Here are 12 most common reasons why a computer turns off without warning:

1. Overheating - (Most popular reason for shutdowns also from Overclocking).
   
2. Hardware Problems ( PSU, CPU, etc)
   
3. Driver Issues
4. Malware Infection
5. Battery Problems
6. UPS / Surge Protector Failure
7. Insufficient Charger Voltage
8. Dodgy Sleep Mode
9. Fast Startup Disadvantages
10. Out-of-date BIOS
11. Software Failures
12. Problematic OS

The answers in Parenthesis was added by me.

[threeheadedsheep]

Understood, I will get back to the forum once the parts have arrived to redo the loop. It's weird because it would reset under no load. I'll have results soon though. 

[elstaci]

Some Users with 3rd generation processors, here at AMD Forums,  have been having the same type of problem where the computer will shut down while on idle with good CPU temperatures.

There is a new AMD BIOS setting concerning the PSU that seems to help with some Users to fix that. Not all BIOS has this new Setting though.

Its called Power Supply Idle Control. In BIOS you need to change it to "TYPICAL" which will fix your Ryzen CPU from shutting down during idle. This worked for some Users but not all.

   `

[Zenpanda87]

Go in bios and first find:

1. Precision boost overdrive and change auto in disable. 

2.  Core performance boost change in -disable

3. XMP- change in enable.       

4. CPU clock ratio- find good and stability :manually find stable processor operation in a bios at a certain frequency. Start with:
40.00
41.00
42.00
etc .....
Test each frequency you raise in windows by setting the CPU load to 100% and raising youtube videos of 4 k resolution (3 vide min) If it is stable, then go back to the bios and raise etc.
I lost a lot of hours with a friend until we realized the essence of the problem.
My motherboard is made of gigabytes.

When you testing amd procesor looking temperature(in procesor) and nuclei(they shouldbe as consistent as possible during operation)
Good luck.Hf.

[h00ti3]

Did you ever find a solution to this problem? My best friend has the same cpu with the same problem. He hasn’t replaced the power supply yet but we’ve been unable to determine the cause of the random crashing.

He is running a 3070ti with 2 x 16GB 3200mhz memory.

I usually OC my builds, so when I was looking at CPU-Z tonight, I noticed the 36x multiplier boosting up to 43x then back down to 36x. This obviously isn’t a problem, and his how the machine should be working, but I usually do a hard overclock and therefore I never see that number move. It’s always locked at 50x on my i7.

So this got me thinking about just hard coding him to 36X in the bios (instead of Auto) and stabilize the processor and just see if that would fix it.

While I was in there, I noticed his Memory was set to Auto as well. I manually set that to 3200 MHz, as per its advertised speed. We played Fortnite and Warzone for a couple of hours tonight with no crashes.

I don’t want to say it’s fixed (knock on wood), but we’re not seeing crashes, at least not for the first couple of hours.

I know this is an old post, and maybe you’ve moved on or that machine is sitting in the corner being constantly glared at, but maybe this is something you hadn’t thought of or tried that might fix it?!