Overheating or Driver conflict?

[roubehrt]

Hey,

 

I talk about following budget pc specs:

 

 

MB: Gigabyte B450 Aorus Elite AMD B450 So.AM4 Dual Channel DDR4 ATX

PSU: 550 Watt Corsair CX Series CX550 Non-Modular 80+

RAM: 16GB G.Skill Aegis DDR4-3200  (Dual-Channel)

Graphics Card: 8GB XFX RX 590 FATBOY OC+ 

 

So basically what happened to me:

 

Everything was fine, i built  the pc myself (beginner). I could play games, even Cyberpunk was running without too much issues and i didnt have crashes.

Didnt overclocked anything myself, took everything as ordered. Overall Temps are quite good.

BIOS seems up to date, PSU i guess should be enough for now.

 

I assumed following made problems:

 

I changed the stock cooler of the RX590 to a Arctic Accelero III (got it from my brother, was barely used he was very happy with it tho).
So basically i saw thats a bit of a tight fit, so i put heatsinks on VRMs and VRAMs. Tho the Cooler covers alot of the VRAMs space i moved the heatsinks of the VRAMs to the outer half (approx . 7 now have heatsinks on half of available space).

1 VRAM didnt have space for a heatsink at all and remained uncovered.

 

I gave it a try and plugged it in. Worked for the first hours and i played games too (for example Battlefield 4 and 5).

Overall GPU temps where (very) fine.

Then i got the driver update of AMD and after it the troubles happened: after some minutes of playing pc started to freeze, short time after i was thrown to my desktop with AMD saying "Driver timeout..."

 

So, i took the Card out and checked Graphics Card, everything was just tight as i put it in before.

Then i removed Drivers via DDU (without safe mode, gave me error messages that it doesnt have acces in safe mode???) and AMD CleanUp.

I installed the driver i had before. But again it continued to crash!

 

Also GPU Fans on 95% made no difference to the crashes.

 

So i assume i did something wrong with the mounting of the cooler right? 

Sadly i cant revert now to stock because i destroeyd the heatpads and need to buy new first.

 

Any advice from you guys? I would very much appreciate it. Thanks!

 

 

[RAS_3885]

Have you tried an older driver version?

 

But as a note, uncovered/partially covered VRMs or VRAM chips would make me nervous. They get HOT, and leaving them with inadequate cooling will be a quick way to fry the card.

[roubehrt]

17 minutes ago, RAS_3885 said:

Have you tried an older driver version?

 

But as a note, uncovered/partially covered VRMs or VRAM chips would make me nervous. They get HOT, and leaving them with inadequate cooling will be a quick way to fry the card.

Yes older version crashes too

 

really? i read somehwere here, VRAMs doesnt need to be fully covered. then i see no way to mount that Accelero on it. Would be interesting tho how there comes a driver timout. What exactly does happen there?

[RAS_3885]

5 minutes ago, roubehrt said:

Yes older version crashes too

 

really? i read somehwere here, VRAMs doesnt need to be fully covered. then i see no way to mount that Accelero on it. Would be interesting tho how there comes a driver timout. What exactly does happen there?

Hmm... maybe VRAM doesn't get as hot as I was thinking then. If you've got info from somewhere that says that's okay for an aftermarket cooler then I'd go with that. I was just speculating.

 

AMD drivers have always been notoriously problem-filled. Other than trying DDU again and reinstalling I'm not sure what else to suggest.

 

Good luck!

[d4n1Xd34th]

5 hours ago, roubehrt said:

Hey,

 

I talk about following budget pc specs:

 

 

MB: Gigabyte B450 Aorus Elite AMD B450 So.AM4 Dual Channel DDR4 ATX

PSU: 550 Watt Corsair CX Series CX550 Non-Modular 80+

RAM: 16GB G.Skill Aegis DDR4-3200  (Dual-Channel)

Graphics Card: 8GB XFX RX 590 FATBOY OC+ 

 

So basically what happened to me:

 

Everything was fine, i built  the pc myself (beginner). I could play games, even Cyberpunk was running without too much issues and i didnt have crashes.

Didnt overclocked anything myself, took everything as ordered. Overall Temps are quite good.

BIOS seems up to date, PSU i guess should be enough for now.

 

I assumed following made problems:

 

I changed the stock cooler of the RX590 to a Arctic Accelero III (got it from my brother, was barely used he was very happy with it tho).
So basically i saw thats a bit of a tight fit, so i put heatsinks on VRMs and VRAMs. Tho the Cooler covers alot of the VRAMs space i moved the heatsinks of the VRAMs to the outer half (approx . 7 now have heatsinks on half of available space).

1 VRAM didnt have space for a heatsink at all and remained uncovered.

 

I gave it a try and plugged it in. Worked for the first hours and i played games too (for example Battlefield 4 and 5).

Overall GPU temps where (very) fine.

Then i got the driver update of AMD and after it the troubles happened: after some minutes of playing pc started to freeze, short time after i was thrown to my desktop with AMD saying "Driver timeout..."

 

So, i took the Card out and checked Graphics Card, everything was just tight as i put it in before.

Then i removed Drivers via DDU (without safe mode, gave me error messages that it doesnt have acces in safe mode???) and AMD CleanUp.

I installed the driver i had before. But again it continued to crash!

 

Also GPU Fans on 95% made no difference to the crashes.

 

So i assume i did something wrong with the mounting of the cooler right? 

Sadly i cant revert now to stock because i destroeyd the heatpads and need to buy new first.

 

Any advice from you guys? I would very much appreciate it. Thanks!

 

 

Same issue, but on nvidia since 2-3 days back

[Mark Kaine]

5 hours ago, RAS_3885 said:

maybe VRAM doesn't get as hot as I was thinking then.

I'm also not an expert in this at all but you're right they get hot, and they need appropriate cooling, not sure if an aftermarket cooler that may or may not covers them is enough. 

 

7 hours ago, roubehrt said:

Then i removed Drivers via DDU (without safe mode, gave me error messages that it doesnt have acces in safe mode???)

well that is definitely odd... I'd try this again, but yeah I'm also not sure. 

 

Last resort would be fresh install of windows I guess, but first you should get some adequate cooling for the vram... 

 

 

1 hour ago, d4n1Xd34th said:

Same issue, but on nvidia since 2-3 days back

I'd make a new thread about this, different specs and all, not to mention different gpu brand... 

 

And guys, *never ever* update a driver that works fine (unless you change gpus I guess) 

every update on any device always bears the risk of breaking something, so why take the risk, as long it works (which is in my experience at least a year...)