Question My Custom Build Is a Nightmare

bonkers_bananas

Prominent
Dec 19, 2020
29
0
530
0
I will be adding minidumps and everything later tomorrow but im pretty much at a loss here i cannot figure out what in the hell i need to do to get my pc stable. here is a brief history on this build...
march of 2019 i built my pc and it consisted of:
ryzen 5 3600x
msi gaming plus max b450 mobo
650w EVGA gold psu
32gbs of 3000mhz ddr4 from OLOY
my old corsair AIO cooler
2x 1tb HDD's from my old build
1x WD m.2 nvme
nvidia zotac 1660s
in the first 2 weeks of this build being functional it would bsod randomly, sometimes it would let me play for days on end sometimes it would crash every 10-20 mins like clock work so the wallet came out and the trouble shooting began, i have changed everysingle component except the cpu, i had used my old 1060 and would get the same crashes i used some spare 2133mhz sticks, same crashes. all signs are pointing to the cpu so after changing all that hardware and doing the following software tests: sfc scan, x86memtest, updated all and every single driver i could, updated windows and pretty much any and all things under the sun that were presented as possible solutions on this forum as well as reddit nothing worked i did however get a couple months of stability by undervolting and underclocking the processor and putting a 85% power limit on the gpu. all was fine until recently.....

the bluescreen has returned but now its even more inconsistent. sometimes the pc wont fire a error code and actually bsod, it will hang until it gets choked out, and applications just randomly close and and fire random error codes that will solve themselves if i just restart the system until an hour or 2 passes and the problem arises again, everything from games to google chrome randomly crash. from the research i've done this is a fairly common issue when it comes to ryzen cpus after trying everything under the sun im ready to just say screw it and swap over to intel and buy and new mobo and cpu. i however do not currently have the income to make this swap so this is a hail mary if anyone has any suggestions of what i can do to fix this issue once and for all im all ears..
 

bonkers_bananas

Prominent
Dec 19, 2020
29
0
530
0
and what i mean by common is it seems this issue persists in mainly amd cpus paired with nvidia gpus which i saw a update was put out to rectify this but after that update fired i did a fresh windows install/ bios update and a chipset driver updates and the issue remained. like i said the only way ive been able to get the build to remain stable for long term use is by undervolting the cpu and limiting the power on the gpu
 

bonkers_bananas

Prominent
Dec 19, 2020
29
0
530
0
If tweaking the power makes it stable, to me, that points to either supply of power (PSU) or power management (Mobo).

Do you have another power supply you could try or mobo?
Yes I swapped from a b450 to a x570 and tried a 750w modular power supply and the problem still persists
 

bonkers_bananas

Prominent
Dec 19, 2020
29
0
530
0
i used my old 1060 and the problem was still persistent and ive changed sata cables but ive never swapped my m.2 to the second port might give that a shot here
 

Amddefector

Prominent
Sep 5, 2020
274
27
720
6
i checked- no bent pins, and i would say it passes the memory test with one stick both times and two sticks
have you monitored all the voltages with something like hw monitor to see I've there's any voltage spikes or over voltage on anything? Different hdmi/dp cable? Other cables, all the simple things. Like all the little risers that the mb screws to in the case, are there only the ones you need in there. No extra ones touching the back of the board?
 

bonkers_bananas

Prominent
Dec 19, 2020
29
0
530
0
have you monitored all the voltages with something like hw monitor to see I've there's any voltage spikes or over voltage on anything? Different hdmi/dp cable? Other cables, all the simple things. Like all the little risers that the mb screws to in the case, are there only the ones you need in there. No extra ones touching the back of the board?
yeah nothing like, that i've had it on my little test bench set up and eliminated the extra hdd's and still bsods but i have noticed in ryzen master that just moments before it will hang/bsod the EDC meter goes into the red and within about 30 secs of that happening it will crash.
 
Exactly what PSU's have you tried with they system? Exactly what motherboard are you currently using? Exactly what kit of RAM are you using? Please list exact model numbers and branding! Undervolting AMD CPUs can cause its own issues. I have personally experienced issues with undervolting with AMD CPUs. Considering the many fresh installs of windows I would conclude that the OS and drivers not the issue. This assumes you actually did a clean and fresh install of Windows and that you do not have parts of the old OS on any drives you are using.
 

Amddefector

Prominent
Sep 5, 2020
274
27
720
6
yeah nothing like, that i've had it on my little test bench set up and eliminated the extra hdd's and still bsods but i have noticed in ryzen master that just moments before it will hang/bsod the EDC meter goes into the red and within about 30 secs of that happening it will crash.
Not many options left then. Apart from trying another cpu I think some bios settings need looking at. First of all you need to use hw monitor, it gives you a lot more information than ryzen master. Second. Resetting the bios to factory settings doesn't necessarily mean they are the right settings! My last am4 board on factory settings was allowing the cpu to run way above its power limits! Using hw monitor you can see 3 values. Lowest, current and highest. So it will give you some idea what the highest voltages, power draw. Temperature etc...

Things like precision boost overdrive is this enabled in the bios? Try disabling it and anything else that boosts the cpu. If you have undervolted the cpu is turbo boost still working? Running the cpu at base clock undervolted is fine but when it boosts its not gonna get the power it needs and run into problems. Can you lock the multiplier so it doesn't boost undervolted. Is there a optimised value setting in the bios? Have you tried that? Before replacing the cpu I would double check all the bios settings.
 

bonkers_bananas

Prominent
Dec 19, 2020
29
0
530
0
Not many options left then. Apart from trying another cpu I think some bios settings need looking at. First of all you need to use hw monitor, it gives you a lot more information than ryzen master. Second. Resetting the bios to factory settings doesn't necessarily mean they are the right settings! My last am4 board on factory settings was allowing the cpu to run way above its power limits! Using hw monitor you can see 3 values. Lowest, current and highest. So it will give you some idea what the highest voltages, power draw. Temperature etc...

Things like precision boost overdrive is this enabled in the bios? Try disabling it and anything else that boosts the cpu. If you have undervolted the cpu is turbo boost still working? Running the cpu at base clock undervolted is fine but when it boosts its not gonna get the power it needs and run into problems. Can you lock the multiplier so it doesn't boost undervolted. Is there a optimised value setting in the bios? Have you tried that? Before replacing the cpu I would double check all the bios settings.
i havent tried to lock the multiplier, i dont have the boost or xmp enabled. so ill give that a shot here to see if i cant produce another minidump file.
 
I personally had a build with 3600G, where I tried to use OLOY memory set which I eventually had to swap for Corsair, was not working properly with the CPU and was giving similar trouble...
- I would check with a pair of sticks that are on the QVL... or at least run memtest64 and pass.
In addition, there was a question a few months ago with a similar problem which ended up being a faulty CPU, so if the above does not help - that would be my next step.
 

bonkers_bananas

Prominent
Dec 19, 2020
29
0
530
0
yeah so on the ryzen 5 mem test passed on both sticks individually and together. so i ran another last night after i put the new cpu in, one stick fired a zillion errors the second i tested it, culprit has been found but i have 0 idea why it wouldn't show in the 1000 memtests i did on the ryzen 5, i let memtest run for 6 hours on the working stick when i checked it before work it had 0 errors so bad ram stick i guess
 

ASK THE COMMUNITY

Latest posts