I dont know how they tested it at the store. I suppose they tried it on a working system and concluded it was the CPU. which was faulty. Or it could be they knew the problem, when I RMA'd I sent them a description of my troubleshooting
So the problem went away for you after you received a new Motherboad and CPU? In that case really you don't know for sure if it was the Motherboard or CPU that was the culprit, other than the store's word for it.
What I am insinuating is that it could have been the motherboard as well, if indeed you had both motherboard and cpu replaced.
I said I received the old board, with a new cpu.. the CPU was defected. I have exactly the same hardware as before, except the CPU is new. And now it works fine.
Appart from the store being an IT store with a very high volume of sales, they are the biggest in the country meaning they have the expertise, I'm also an electronics engineer, with a lot of experience in assembling IT hardware, and I'm a software developer in Linux. I have over 10 years experience in all this. Believe me the CPU was defected. Its not so abnormal you know, new line of CPUs, QC is probably failing. Eventually they'll fix it I'm sure. They have to, to make profit and conquer server market which is where the money is
I have a Ryzen 5 3600, water cooled custom loop (never overclocked other than what Core Performance Boost does), in a
Asus x570-I MB with
32gb of Patriot 3200mhz ddr4 (PVS432G320C6K)
Also I've tested with 16gb of G.Skill 3200mhz ddr4 (F4-3200c14d-16gfx on qvl for MB)
Corsair tx750 PSU
GTX 1080 (also tested GTX 1070 and a different brand of GTX 1080) output via HDMI to a 4k TV
NVME 4.0 boot drive, WD black sata drive,
Windows 10 64bit 2004
Latest Bios as of Sept/2020
D.O.C.P. standard on
Core Performance Boost - auto
Precision Boost Overdrive - Disabled
If I use bios DOCP to run the memory at the settings I paid for, the system will randomly reboot during idle. Typically after windows is just logging in. Also I've had a few random reboots during table top simulator sessions with the memory set to DOCP as well but its much less frequent than during idle.
Changing memory kits did not fix.
If I run the memory on Auto instead of DOCP, it lowers the speed from 3200mhz to 2166 in the case of the Patriot memory and 2400 for the g.skill, the random reboots are gone.
I believe I have a bad CPU as well and have started the RMA process.
I also have a Ryzen 9 3900x on a
Asus x370 crosshair vi hero MB bios 7704 with
G.skill f4-3200c14d-16gfx memory, slot 2/4 populated.
I've don't have the random reboots on this system. I've had some issues and initially I solved them by turning off both PBO and CPB as well as turning off the DOCP.
Currently I have PBO off and CPB and DOCP on auto/standard and I've been stable so far.
Your PSU is over 10 years old and the new RM series are more efficient, modular and have excellent lower power efficiency. The RM750 is an excellent choice.
I ran the 3600 with a new Corsair SF750, as well so I don't think its the PSU. I changed cases so I switched back to a spare full atx psu I had.
I do like the RM, HX, AX series psu's from corsair but they are quite expensive at the moment. If I secure one, I'll test to see if it makes any difference but reading what rui_ddit said, he fixed it by replacing the cpu.
Either way, I'll update with new info in the future.
The Corsair SF line of ITX power supplies are fine. Some issues with some lots of the SF750 were noted a while ago,
We have recently identified higher-than normal RMA rates among our SF family of small-form-factor PSUs. Following a thorough investigation, we have found a potential issue that can manifest when the PSU is exposed to a combination of both high temperatures, and high humidity. This regrettably can cause the PSU to fail. This issue potentially affects units in lot codes 194448xx to 201148xx, manufactured between October 2019 and March 2020.
The lot is the first 6 digits of the serial number, so check it in case you have one of the problematic ones.