Ryzen 3600/B450/GTX1050 Ti system shutting down on game/DX launch to a state requiring PSU switch off/on – fault in GPU?

[Porco Malo]

First of all, I'm a long time listener, first time caller.

I've tried to find similar cases in several places, but haven't figured out how to pin down what is wrong in my system.

 

This is a few months old build, where I've had the same problem come up several times a week.

 

Symptoms:

• Everything else works fine, but on launching PUBG (which I almost exclusively play), about 50-70% of times the PC powers down before reaching the lobby screen (no BSOD, just power off).
• It shuts down to a state where the power button is unresponsive, while RAM is still appears to be powered (RGB on RAM is still on like in sleep mode), and I have to switch the PSU off and on to be able to boot up the system.
• System has shut down similarly upon launching the OpenGL test on Cinebench R15 (which lead me to suspect it could be a GPU issue).
• A few times (less than 1/20) the system has similarly powered down right after booting, when displaying the logon screen.
• Sometimes, less than once in two hours, the screen flickers for a few seconds in 2D/desktop use

 

Also:

• This typically happens 2-4 times in a row after which the game has reset to default settings (graphics settings all the way down to low, key mappings to default etc.)
• I have been in contact with PUBG support and as per their instructions reinstalled the game, reinstalled the GPU drivers, and done a Windows services fault seeking by switching all the non-Microsoft services off and on with MSCONFIG (I was unable to reproduce the issue with this, which again led me to believe this could be HW related).
• When the game successfully launches, it very rarely crashes during gameplay (either with GPU stock clocks or overclocked to core clock +150 MHz & memory +450 MHz) – the in-game performance is just fine
• I have tried repeatedly launching PUBG with GPU under-clocked (core clock & mem) – this doesn't make a difference
• I used the same GPU for almost a year in a Dell Optiplex, without this kind of a problem

 

I don't really play any other games with this system so I can't say if launching some other similar game would be different.

 

I would greatly appreciate any help in fault seeking, I unfortunately don't have any spare/replacement components that I could have used to pin down a possible HW issue. I've exhausted all my own skills

Thank you in advance to anyone taking the time to read this.

 

System:

OS: Windows 10 Pro 64-bit
CPU: Ryzen 5 3600 (stock) – idle 39° C, load 54° C
MB: MSI B450 Tomahawk Max (BIOS 3.60 4/22/2020)
RAM: Corsair Vengeance Black RGB Pro 3600MHz 2x8GB (default XMP setting)
GPU: Gigabyte GTX 1050 Ti LP – idle 30° C, load 73–75° C
SSD: Kingston A2000 M.2 500 GB
PSU: Corsair CV450 450W

Cooler: Cooler Master Hyper 212 Black
Case: Kolink Observatory Lite Mesh RGB

Display: HP Z23n 1920*1080 (overclocked from 60 Hz to 69 Hz) over DisplayPort

[cr8tor]

What you describe makes me think power supply. Any chance you can swap that thing out? I know you dont have parts, a friend maybe? Short term?

[Mark Kaine]

Well there isn't much you can do than starting to exchange parts. 

 

PSU, yes why not, always a good start for trouble shooting these kind of issues. 

 

Next, if issue persists, motherboard. 

 

This specific motherboard is highly suspect in my opinion and I am not in the least surprised it likes to shut down for no reason (had the same issue with it, and a plethora of posts here and elsewhere about similar issues, with this specific board, I think only gigabyte "aorus" might be worse... 

 

 

I'm not saying that's the culprit but it's worth checking for sure after power supply. 

 

And if neither of these things help, well than post again we'll somehow figure out what to do next then. 

 

 

 

ps: first things first tho: did you reset CMOS? 

[Porco Malo]

Thank you so much guys. I cleared CMOS now via jumper but the issue remains.

As one extra detail the issue seems to arise with close to 100% certainty in cases where I wake up the PC from sleep - seems that if it has been switched all the way off, the chance is much lower. Could this in fact point to something in the PSU?

 

At least wattage wise I thought I would be very well covered with the current PSU but I guess I will just have to buy another one and test the system with that. I don't really have anyone nearby to loan one from.

 

God I wish it's not the motherboard, that's gonna be a real warranty hassle, this one is a customer return item, sold 10 months ago...

I actually got this one because I thought it was one of the better ones, I had no idea a lot of folks had problems with it.

[cr8tor]

11 hours ago, Porco Malo said:

Could this in fact point to something in the PSU?

PSU is more likely than mobo from what you are describing.

 

However btw, there is a newer bios.

7C02v37 - Release Date: 2020-06-15 - MSI support page link

Might be worth it.

[Porco Malo]

On 10/27/2020 at 6:06 PM, cr8tor said:

PSU is more likely than mobo from what you are describing.

 

However btw, there is a newer bios.

7C02v37 - Release Date: 2020-06-15 - MSI support page link

Might be worth it.

That's true. I'll look into that too.

 

One more thing I thought of - my GPU doesn't have a separate power connector, so it's just powered through the PCI-E slot. If this crashing would be PSU related, in a situation where power use sharply goes up, wouldn't it then just as well happen when I for example launch the CPU test on Cinebench or encode video etc (I've done both multiple times)?

 

I mean, isn't it all the same kind of load for the PSU, when it just feeds the motherboard (not for example different rails feeding M/B and GPU)?

[cr8tor]

On 10/30/2020 at 8:28 AM, Porco Malo said:

I mean, isn't it all the same kind of load for the PSU, when it just feeds the motherboard (not for example different rails feeding M/B and GPU)?

Your CPU is a 65w cpu and your GPU is a 75w GPU.

Power is power regardless of where it comes from, but your GPU is pulling more than your CPU.
GPU's also tend to ramp up and down a bit more as their idle state tends to be a bit more idle than a CPU's idle state. 

And once you use the GPU it often goes from idle to max so it fluctuates a lot.

This is harder on the PSU and when you add in the CPU is also possibly drawing more power at the same time, the PSU may just not be handling the peak power draw.

[Porco Malo]

On 11/6/2020 at 1:48 PM, cr8tor said:

Your CPU is a 65w cpu and your GPU is a 75w GPU.

Power is power regardless of where it comes from, but your GPU is pulling more than your CPU.
GPU's also tend to ramp up and down a bit more as their idle state tends to be a bit more idle than a CPU's idle state. 

And once you use the GPU it often goes from idle to max so it fluctuates a lot.

This is harder on the PSU and when you add in the CPU is also possibly drawing more power at the same time, the PSU may just not be handling the peak power draw.

This took quite a while, I ordered a Seasonic Focus PSU to use for testing, and the first order got canceled for some reason – now I finally got it, and looks like the fault was indeed in the Corsair PSU. All the issues are gone now. Really happy I don't have to start swapping the motherboard!

 

Thanks again for the help!

 

There's the PSU tier list thread here – do you know how the data ends up on that list? I feel like I should report somewhere the fault with the other PSU (other than RMA'ing it to where I bought it).

[cr8tor]

Good to hear this got resolved. Thanks for sharing!

[cr8tor]

On 12/7/2020 at 12:41 AM, Porco Malo said:

I feel like I should report somewhere the fault with the other PSU (other than RMA'ing it to where I bought it).

You could post on the thread for the tier list, but ultimately, all manufactures have bad parts now and then.

Still worth sharing though.

That or even create a post on the power supply branch of the forum.