SAPPHIRE Pulse Radeon RX 580 screen goes black

[Yuanlee]

Good morning, my name is Yuanlee and you gentlemen are my last hope.

You see, my gpu has been failing since a couple of days ago. After 10 minutes or so of playing games that require a considerable amount of power such as doom eternal or shadow warrior 2 the screen goes black. Everything else still working for 5 seconds or so, but eventually it will freeze and need to reboot.

This is not the first time this happens. Last summer I had the exact same issue, and after trying a lot of things, turned out the issue was that the thermal pads of the gpu were in a very bad state, so after remplacing them with new ones the gpu worked nicely. Thats why I supposed that remplacing the pads would be the solution, but it wasn't.

I'm studying computing, so i asked about this to my teacher and he told me about how this forum gets to the solution very frequently, and here i am, hoping you distinguished gentlemen can help me.

I will attach the gpu-z profile, but you can ask for any aditional information you need, i will be reading.

PS: English isn't my mother tongue, so sorry if i wrote something wrong .

 

[Palindrome]

Hi, one thing that helps with these types of issues is putting your complete system specs either in your profile or in the post.
From what you're describing, I have a few ideas:
1) Make sure the fans actually turn on. Sounds dumb, but there's cases of certain cards having their fans essentially "bug out" and never go out of 0-RPM Mode, although I doubt that'll be the issue
2) What are the temperatures like under load? With this card in particular you wanna aim for low-mid 70s, the card starts to throttle once you get to around 80°, relating to 1) your card may be overheating, specifically as you mentioned it happening last summer (when ambient temps are higher).
3) (due to missing specs) Is your power supply powerful enough? Maybe try a different 8-pin lead, see if that changes anything.
4) With AMD's piss-poor drivers, those being the issue also cannot be ruled out. Using DDU to completely wipe your current driver installation and installing fresh ones straight from AMDs website would be another thing I would try.
5) maybe try reseating the card. take it out of the PCIe slot, use a bit of compressed air or sth to make sure theres no dirt or dust in the slot itself, maybe gently wipe the gold connectors on the card with a microfibre cloth. again, I doubt that'll be the issue, but it's a simple thing to try, just to make sure.

 

[Yuanlee]

Thank you for answering.
My PC specs are:
- OS Windows 10
- 16gb ddr4 ram
- Processor Ryzen 5 1500x
- Motherboard Asus prime b350m-a
- Power supply Nox Urano VX750W
So answering your ideas:
1) Its on. In fact, when the screen goes black the fans start spinning like crazy.
2) Under load the gpu gets to the 80º mark.
3) It should be enough, and already tried a different 8-pin but doesn't work.
4 and 5) I already did, saw no results.

 

[Palindrome]

Alright, thanks for filling those in, the only thing that stands out to me is that the power supply is from a brand I have never heard of, but that doesn't necessarily mean it's bad. When your issue occurs, your screen goes black, GPU fans go to max RPM, and the system is just frozen, meaning you need to manually restart it, or do you get a bluescreen and associated error code?
1) what happens if you run a purely synthetic benchmark, f.ex Furmark at FHD with 8x MSAA (that'll already be enough to load the GPU up fully). Are you able to replicate the issue that way?
2) Have you tried enabling/disabling XMP? I have heard stories of people having all sorts of weird issues with XMP and/or high-er memory speeds on Zen 1, personally I had no problems with a R5 1400 and 3200 speed RAM, but it may be worth toggling it on/off just to see if it makes a difference.
3) The RX580 Pulse is a dual bios card (I used to have one ), have you tried toggling to the other vbios to see whether the issue persists?
4) Is your motherboard's BIOS fully up to date? Bit of a wild-card guess, I'd be surprised if it is the mobo, but then again the entire communication between the GPU and the rest of the system occurs over the PCIe bus, perhaps a newer version of the BIOS resolves an issue of the type that you're describing?
5) Considering you already DDU'd and reseated your card I'm sure you've done this as well, but have you tried clearing the CMOS? And/or replacing the CMOS battery itself, although I doubt it'll be flat after just 5-ish years, those button cells are dirt cheap and widely available so it's an easy thing to try.

 

[Yuanlee]

Hi again

I don't get a blue screen, just lose signal, so the screen turns black, the fans go wild and then the system freeze (I know it freezes because of the audio).

1) I can't replicate the issue, it only happens when I play a videogame.
2) I had it disabled, so i enabled it and try with no success.
3) Mine isn't dual bios actually.
4) It is.
5) I did, but still gets black screen.

 

[Tetras]

I would strongly recommend you use much older drivers for this graphics card. I use 19.11.3 for my Polaris cards. The newer drivers (20 onwards) result in crashes and black screens.

 

[Yuanlee]

I would strongly recommend you use much older drivers for this graphics card. I use 19.11.3 for my Polaris cards. The newer drivers (20 onwards) result in crashes and black screens.

Thank you for answering .

I installed the drivers you sugested, but unfortunately it didn't solve the issue.

 

[droid-I]

So the motherboard BIOS is likely on version 5606?

A couple of things to try, remove one of the RAM sticks, either one in turn, does the issue still persist? (assuming you have 2x8GB in there)

Could also try with another HDMI cable, but that does not explain why it would freeze.

 

[Yuanlee]

So the motherboard BIOS is likely on version 5606?

A couple of things to try, remove one of the RAM sticks, either one in turn, does the issue still persist? (assuming you have 2x8GB in there)

Could also try with another HDMI cable, but that does not explain why it would freeze.

The issue still persist unfortunately. Also it happens with others HDMI cables as well.

I have an idea of what could be the reason though. I did change the thermal pads of the gpu, but i used the pads that were left after the first time i changed it. The first time i bought a pack of 3 bags with 3 pads in each one: 1 0'5mm pad, 1 1mm pad and 1 1'5mm pad. After the remplacement, there were only the 1 and 1'5mm left, so i used them this time to remplace it.

As I have recently discovered, you can't use pads that thick in this gpu because it loses contact with the heatsink, so I just bought a good thermal pad of 0,5mm and I'll remplace it again.

Wish me luck gentleman, Elden Ring is in the horizon and I surely can't afford a new gpu .

 

[eidairaman1]

Alright, thanks for filling those in, the only thing that stands out to me is that the power supply is from a brand I have never heard of, but that doesn't necessarily mean it's bad. When your issue occurs, your screen goes black, GPU fans go to max RPM, and the system is just frozen, meaning you need to manually restart it, or do you get a bluescreen and associated error code?
1) what happens if you run a purely synthetic benchmark, f.ex Furmark at FHD with 8x MSAA (that'll already be enough to load the GPU up fully). Are you able to replicate the issue that way?
2) Have you tried enabling/disabling XMP? I have heard stories of people having all sorts of weird issues with XMP and/or high-er memory speeds on Zen 1, personally I had no problems with a R5 1400 and 3200 speed RAM, but it may be worth toggling it on/off just to see if it makes a difference.
3) The RX580 Pulse is a dual bios card (I used to have one ), have you tried toggling to the other vbios to see whether the issue persists?
4) Is your motherboard's BIOS fully up to date? Bit of a wild-card guess, I'd be surprised if it is the mobo, but then again the entire communication between the GPU and the rest of the system occurs over the PCIe bus, perhaps a newer version of the BIOS resolves an issue of the type that you're describing?
5) Considering you already DDU'd and reseated your card I'm sure you've done this as well, but have you tried clearing the CMOS? And/or replacing the CMOS battery itself, although I doubt it'll be flat after just 5-ish years, those button cells are dirt cheap and widely available so it's an easy thing to try.

https://www.reddit.com/r/buildapc/comments/f47r82
I'd avoid them like the plague.

I would strongly recommend you use much older drivers for this graphics card. I use 19.11.3 for my Polaris cards. The newer drivers (20 onwards) result in crashes and black screens.

Use Radeon Pro for Enterprise

The issue still persist unfortunately. Also it happens with others HDMI cables as well.

I have an idea of what could be the reason though. I did change the thermal pads of the gpu, but i used the pads that were left after the first time i changed it. The first time i bought a pack of 3 bags with 3 pads in each one: 1 0'5mm pad, 1 1mm pad and 1 1'5mm pad. After the remplacement, there were only the 1 and 1'5mm left, so i used them this time to remplace it.

As I have recently discovered, you can't use pads that thick in this gpu because it loses contact with the heatsink, so I just bought a good thermal pad of 0,5mm and I'll remplace it again.

Wish me luck gentleman, Elden Ring is in the horizon and I surely can't afford a new gpu .

Get hwinfo64 and get screenshots of temperatures, volts.

 

[Yuanlee]

Good news gentlemen!

Turns out it was in fact the 1mm thermal pad, I remplaced all of them with a 0'5mm and now the gpu works nicely!

Thank you all for helping me out with this issue, now I can enjoy Elden Ring this friday!

I'd avoid them like the plague.

Thank you eidairaman, i had no idea this power supply was so dangerous! I bought it because a good friend recommended it, but I'll buy a new one given the circumstances.

And again, thank you all for helping me out, love you guys .

 

[de.das.dude]

https://www.reddit.com/r/buildapc/comments/f47r82
I'd avoid them like the plague.



Use Radeon Pro for Enterprise




Get hwinfo64 and get screenshots of temperatures, volts.

i was about to say... that PSU is very suspect.

Dont listen to the others saying drivers this and that.

I have never faced any driver issues with AMD since the last 10 years. The reason i upgraded was because my older gpu is so old, it doesnt get any drivers hahaha. ANd i never had to change thermal pads.
just make sure to remove them carefully, i clean them with some IPA and put a touch of TIM before pressing down.

glad it was a different issue and the PSU did not kill your system

 

[Otiic]

Good news gentlemen!

Turns out it was in fact the 1mm thermal pad, I remplaced all of them with a 0'5mm and now the gpu works nicely!

Thank you all for helping me out with this issue, now I can enjoy Elden Ring this friday!



Thank you eidairaman, i had no idea this power supply was so dangerous! I bought it because a good friend recommended it, but I'll buy a new one given the circumstances.

And again, thank you all for helping me out, love you guys .

Well, I've been having almost the same issue than yours, last summer I did too but it went away by itself, now this summer I'm having the same problem, but even worse than last time, I've tried everything, so the only thing is left to do is this, but I don't know how to do it, so I'm taking my pc this week to a confident place to do it for me, and they will do a maintenance for all the pc too. Hope it fixes the problem, wish me luck.