GPU randomly crashes and screen goes black?! Radeon RX 580

[Miheyu]

Let me just start off my saying I know... This is what I get for buying a used card off eBay, I thought getting a Radeon RX 580 Armor for $120 was a steal, but maybe I should have known better. My whole rig has been built on sales and budget parts so I guess it's par for the course. This is the life for a broke ass gamer/streamer/struggling artist like myself with a wife and 3 kids to priorities... "But I digress!"

 

I recently upgraded to a Ryzen 5 2600 build on an Asus B450 motherboard. I've got 2 sticks of G.Skills Aegis 8gb DDR4-3000 and I was using a GeForce GTX 660ti that I got from a friend. After watching tirelessly for sales I finally decided to browse around eBay and found a few Radeon cards, I sent this guy an offer of $120 + $9.95 and he accepted. It was listed as Lightly Used purchased in late 2017. I got it home, upgraded the computer and it worked great for the first few couple of days. Then I started to get unprompted black screens and the fans on the GPU would ramp up full blast. It would happen after games, and one time while I was watching a YouTube video of Linus. The screen went black but I could still hear the video playing in my headset. I had HWMonitor installed and tried to keep it running in the background, but it never seemed like the temps got crazy. During games I would see it climb up to about 80c, but I wasn't watching it that close when I was on YouTube so I can't say what it was doing. 

 

Before you link me to the thread about the GPU crash (I've read through it), YES I did go get fresh Thermal Compound and I've replaced the crunchy junk on the GPU. I also installed MSI Afterburner so that I can keep better tabs on my temps. I actually played around with the custom fan curve so that I always have a small amount of idle fan cooling. So far, things seem to be looking good. I had idle temps around 52c-55c and after setting up a custom fan curve is looks like I'm at an idle of 41c. 

 

I guess the reason I'm posting here is because I thought you guys could help me figure out what's going on with this thing. Should I be doing anything else? I can't afford a brand new card, and this budget build has been a labor of love, but I worry that I'm just prolonging the inevitable. This is what I get for venturing into the seething underbelly of used Graphics Cards! Thank you in advanced for any help and suggestions!! Wish me luck.

[DailyProcrastinator]

What PSU are you using and how old is it?

[Miheyu]

I'm using a Thermatake Smart 700w PSU, It's only like 3 years old, maybe? If the power was lacking would the fans on the GPU still work and run? The card turns into a jet engine and I can't get a picture back on the screen

[DailyProcrastinator]

13 minutes ago, Miheyu said:

I'm using a Thermatake Smart 700w PSU, It's only like 3 years old, maybe? If the power was lacking would the fans on the GPU still work and run? The card turns into a jet engine and I can't get a picture back on the screen

Welcome to the forum!

 

You need to 'Reply' to comments in order for people to see them, this can be done by clicking the arrow icon at the bottom of a comment. Otherwise, people will not be notified of your responses.

 

Anyways, others may read into this differently, and I am certainly just one opinion but that is quite frankly a terrible quality PSU. Voltage ripple can cause system crashes, this has nothing to do with being 700W because while it can sufficiently power your system (a 450W PSU would do fine), it is made with cheap components that under load can deliver the power in a way that can harm components (kill them even) or even cause systems crashes. But before you go off and get a new PSU there are many other items to consider. I only bring up a faulty PSU as someone on here once was in a similar situation and it indeed ended up being the PSU.

 

Although it is not likely to see system crashes, errors in your GPU drivers could cause issues, you should re-install your GPU drivers but remove the old ones first using DDU. That is step one.

 

Are you OC'ing your CPU or GPU? What speed is your RAM running at?

[Miheyu]

4 minutes ago, JM21 said:

Welcome to the forum!

 

You need to 'Reply' to comments in order for people to see them, this can be done by clicking the arrow icon at the bottom of a comment. Otherwise, people will not be notified of your responses.

 

Anyways, others may read into this differently, and I am certainly just one opinion but that is quite frankly a terrible quality PSU. Voltage ripple can cause system crashes, this has nothing to do with being 700W because while it can sufficiently power your system (a 450W PSU would do fine), it is made with cheap components that under load can deliver the power in a way that can harm components (kill them even) or even cause systems crashes. But before you go off and get a new PSU there are many other items to consider. I only bring up a faulty PSU as someone on here once was in a similar situation and it indeed ended up being the PSU.

 

Although it is not likely to see system crashes, errors in your GPU drivers could cause issues, you should re-install your GPU drivers but remove the old ones first using DDU. That is step one.

 

Are you OC'ing your CPU or GPU? What speed is your RAM running at?

If by Reply you mean "Quote" then I shall try. Sorry, new to the forums as you noticed. Thank you for the info on Voltage Ripple, I had no idea about any of that. 

 

No, I'm not overclocking anything. Don't tell anyone, but I'm still using the Wraith Cooler on my Ryzen 5. I was hoping that the upgrade would be enough for me, so I hadn't messed with OC'ng the GPU. Looks like I'm running the RAM at 2133... hmmm, need to fix that in BIOS I guess?

[DailyProcrastinator]

Just now, Miheyu said:

If by Reply you mean "Quote" then I shall try. Sorry, new to the forums as you noticed. Thank you for the info on Voltage Ripple, I had no idea about any of that. 

 

No, I'm not overclocking anything. Don't tell anyone, but I'm still using the Wraith Cooler on my Ryzen 5. I was hoping that the upgrade would be enough for me, so I hadn't messed with OC'ng the GPU. Looks like I'm running the RAM at 2133... hmmm, need to fix that in BIOS I guess?

Yes, 'Quote', and no worries.

 

The Ryzen stock cooler is fine, I would run everything at stock settings until your issue is rectified. But as I stated install DDU, remove all old GPU drivers, including the old GTX 660 Ti Nvidia drivers. Then re-install new Radeon drivers. That is step one.

 

Does the used card have any kind of buyers protection? It is possible that the card was defective before you received it. 

[Miheyu]

33 minutes ago, JM21 said:

Yes, 'Quote', and no worries.

 

The Ryzen stock cooler is fine, I would run everything at stock settings until your issue is rectified. But as I stated install DDU, remove all old GPU drivers, including the old GTX 660 Ti Nvidia drivers. Then re-install new Radeon drivers. That is step one.

 

Does the used card have any kind of buyers protection? It is possible that the card was defective before you received it. 

Thanks so much for the help! I'm reinstalling Radeon drivers now. 

 

It's entirely possible that there was something wrong with the card! That's why I regret buying on eBay now, but it was a hell of a deal (and I'm so broke that it was my only option). There is an option to return, if push comes to shove. I just hope there is something I can fix, within reason, before I send it back. 

[DailyProcrastinator]

8 minutes ago, Miheyu said:

Thanks so much for the help! I'm reinstalling Radeon drivers now. 

 

It's entirely possible that there was something wrong with the card! That's why I regret buying on eBay now, but it was a hell of a deal (and I'm so broke that it was my only option). There is an option to return, if push comes to shove. I just hope there is something I can fix, within reason, before I send it back. 

No problem! Yeah GPU being defective is very possible. 

[Fasauceome]

1 hour ago, JM21 said:

Although it is not likely to see system crashes, errors in your GPU drivers could cause issues, you should re-install your GPU drivers but remove the old ones first using DDU. That is step one.

Yup, this should be step 1.

[Miheyu]

Well... I thought things were looking up for me, but I guess that's too much to ask for. Temps were stable around 42-46c and I got all the drivers uninstalled and then re-installed. I left my computer on and went to bed...

 

When I got up in the morning, I sit down in front of a black screen and the sounds of GPU fans blasting. I'm requesting a refund, and I've got the old 660 ti back in the rig. Thanks for all the help guys