[SOLVED] GTX 970 Black Screen Crash - dying GPU?

[EJWellwood]

System specs:
GTX 970 (palit)
16GB RAM (can't even remember brand, but it's cheap (and bad))
AMD 9590 (yes, I know, silly)
Corsair TX750M PSU
250GB SSD
Windows 7

For a while now I've been getting crashes whenever I'm gaming - more demanding games will lead to more frequent crashes. I can play for a few minutes and then I get a black screen and the sound plays in the background for a few seconds then goes crazy and then I'm forced to hard-restart (although sometimes it does restart by itself). I have been using a back-up card instead (GTX 660ti) but it really can't handle much these days so I want to try and fix my 970. I really don't know what's causing this issue: a driver issue (reinstalled drivers multiple times with custom installation); temperature (it's not getting hot); power (just got a new PSU that should handle the load). It's only under high stress that it will crash, never crashes when browsing internet etc.

I've had a lot of problems with my pc the last two years and haven't had the money or patience to fix them, now that Corona is around I want to finally fix them. This issue must have started around this time last year after my HDD packed in and I shifted over to my SSD. It's weird as before that my pc was running fine, never any issues with the 970. My PSU also died a few months later after playing Battlefield 4 - a loud pop and a nasty smell = rip. I've recently acquired a new PSU hoping that the GPU issues would be fixed, but they haven't and I'm really hoping that the graphics card isn't dead 🙁

If I'm honest I don't even know if it's the GPU that's causing the issues, it could very well be the CPU. I don't know if shifting over to my SSD and reverting from Windows 10 to Windows 7 might have seriously messed up drivers too.

I have not tested the GPU with other rigs, and probably won't be able to.

EDIT: I don't know why I forgot to mention the fact that it doesn't ALWAYS fully crash. I can still usually control my computer, e.g. I can pause and play videos and still move around in-game and sound plays etc., but the screen has gone black (crashed GPU) and reconnecting HDMI does nothing, leading to me having to restart.

Anyways, help is much appreciated - thank you!

 

[Barty1884]

What motherboard do you have? How are you cooling the 9590?

 

[EJWellwood]

What motherboard do you have? How are you cooling the 9590?

The motherboard is an ASUS Sabertooth 990FX

I used to have cooling issues with the processor but ever since I put a Noctua NH-D15 in it's been fine

 

[Barty1884]

The motherboard is an ASUS Sabertooth 990FX

I used to have cooling issues with the processor but ever since I put a Noctua NH-D15 in it's been fine

It's not so much temperatures I'm thinking about there (although that's a good consideration), it's the VRMs.
The 9590 is a ridiculous CPU, with crazy power requirements. Problem could be in the power delivery; VRMs overheating, or degraded etc. An option could be to point a case fan (or even a desk fan for the purposes of troubleshooting) at the VRM/socket area.... a little tough to do with an NH-D15 in the way though.

The 9590s, even with a fairly 'basic' issue, can be a nightmare to troubleshoot.

A GPU driver wipe (with DDU in safemode) and reinstall of the latest is an option, as is a disassembly & clean of the GPU. I'd be surprised if the 970 is presenting issues before the 9590 &/or motherboard is though...

 

[EJWellwood]

It's not so much temperatures I'm thinking about there (although that's a good consideration), it's the VRMs.
The 9590 is a ridiculous CPU, with crazy power requirements. Problem could be in the power delivery; VRMs overheating, or degraded etc. An option could be to point a case fan (or even a desk fan for the purposes of troubleshooting) at the VRM/socket area.... a little tough to do with an NH-D15 in the way though.

The 9590s, even with a fairly 'basic' issue, can be a nightmare to troubleshoot.

A GPU driver wipe (with DDU in safemode) and reinstall of the latest is an option, as is a disassembly & clean of the GPU. I'd be surprised if the 970 is presenting issues before the 9590 &/or motherboard is though...

I tried the driver wipe, gave everything a good clean and set up a fan - it didn't work. Thinking on it I think you're right with the VRMs as it did seem to last a little longer before crashing after doing that stuff. I regret getting this processor (stupid logic of young me thinking bigger = better) and it looks as though I'm gonna need a new CPU - a bit cheeky, but do you have any recommendations?

Thanks for helping me anyways, made things a lot easier and saved me from getting a new GPU that probably would have made things worse (I'm presuming it's okay to use the 660ti for the moment without further damaging things?)

 

[Barty1884]

Do you have access to another system that you could test the 970 in? If you can replicate the workload on the 970 elsewhere and it doesn't crash... then you have an issue within your specific hardware config, and the 9590 or MB is the most likely culprit.

As far as new recommendations go, what kind of budget would you be working with? And where are you located?

 

[EJWellwood]

Do you have access to another system that you could test the 970 in? If you can replicate the workload on the 970 elsewhere and it doesn't crash... then you have an issue within your specific hardware config, and the 9590 or MB is the most likely culprit.

As far as new recommendations go, what kind of budget would you be working with? And where are you located?

I don't have another system to test things in - nor can I really test in a friend's either given the current circumstances.

I would like to be spending less than £200, but that's dependent on whether or not I have to replace the mobo or not. I might just wait a while for my student finance to come through first before committing to a big spend - thanks for your continued help!

 

[Barty1884]

Makes it a little tough to definitively say then, but I'd suspect chances are pretty good.

200quid +/- isn't bad ballpark.... even including a MB/DDR4.

As an example:

PCPartPicker Part List

CPU: AMD Ryzen 5 1600 (14nm) 3.2 GHz 6-Core Processor (£94.97 @ Laptops Direct)
Motherboard: Gigabyte B450M DS3H Micro ATX AM4 Motherboard (£65.47 @ Ebuyer)
Memory: Crucial Ballistix 16 GB (2 x 8 GB) DDR4-3000 Memory (£69.98 @ Amazon UK)
Total: £230.42
Prices include shipping, taxes, and discounts when available
Generated by PCPartPicker 2020-04-09 21:12 BST+0100

Now, officially that's a Ryzen5 1600 first Gen, but those are being phased out and there's a pretty good chance you'd get an "AF" product sku, which is a Ryzen5 2600 is all but name.

 

[BelgianWarGod]

I don't know why I forgot to mention the fact that it doesn't ALWAYS fully crash. I can still usually control my computer, e.g. I can pause and play videos and still move around in-game and sound plays etc., but the screen has gone black (crashed GPU) and reconnecting HDMI does nothing, leading to me having to restart.

If you get a black screen try this combo: win+ctrl+shift+b this will restart the gpu driver. If you're screen comes back to life you know it's a bad driver. Something tells me it's not a bad mobo or gpu, maybe win10 => win7 caused a "bug".

 

[EJWellwood]

I think Windows 10 has the win+cntrl+shift+b feature but not Windows 7. I tried to get it to crash whilst hovering over the device manager but it's not really possible. Plus it seems to crash when engaging in CPU intensive stuff/loading in, e.g. playing a battle in Bannerlord is fine, it's just as soon as I loadout from it onto the campaign map it crashes (for other games there's different triggers). I've had this CPU for quite a while now so I guess several years of being quite toasty have probably taken its toll.

Thanks anyways for the suggestion Belgian, going from 10 to 7 did seem to mess quite a lot of other things up too so it's no surprise it's playing some part in everything.

This list seems great too Barty - it's probably time I upgrade some things!