[SOLVED] My GPUs won't output display.

Feb 14, 2022
8
0
20
Hey. Long time browser and I usually get most of my info on troubleshooting when I'm stuck from this website, but I've encountered a very serious issue that I need a lot of help with. Normally I can deal with these things myself, but this has me completely stumped and I'm out of ideas.

A few days back, my GPU (an RTX 2070 Super) was under relatively heavy load gaming for maybe an hour, and the computer just completely shut down in the middle of use. Since then, the PC would start, but there was no display, so I took it apart to see how it was inside when it was powered on. The GPU powers on with the fans at 100% and the usual RGB lighting, but still won't display anything.

At one point, smoke came out of the machine and I couldn't find what the cause of it, but apparently the Power LED+ cable may have been the culprit since it disconnected from it's plastic connector. This doesn't seem to have any effect powering the machine on, so I've just left it severed and disconnected the Power LED- cable from the motherboard too. I can't find any signs of anything burnt on the motherboard either (maybe I accidentally knocked the Power LED+ cable out of its socket, and it burnt up that way).

Anyway, I have a spare GPU (an RX Vega 64) and that works in very short bursts. I can sometimes get to the BIOS if I keep rebooting, but the monitor turns black again shortly after reaching it, so I can't even access it to do anything. For a little while, it got as far as the Windows login screen before dying, but it hasn't reached that since two days ago. I've tested the monitor by plugging a laptop into it, and it works fine.

I thought the PCI-E slot was the problem, so I switched the card to the other one (the x8 slot), but I still get the exact same issue with both cards (RTX2070 fans spin 100% with no output; Vega 64 only sometimes giving me very short output, and nothing otherwise). Also, I've tried HDMI and Displayport, and still have the same issue.

I don't have onboard graphics either, so I can't test that out, and I also can't test the cards out on other machines since I don't have access to any.

Here are my specs:

ASUS X470 ROG Crosshair VII Hero
Ryzen 3700x
ASUS ROG Strix GeForce RTX 2070 Super OC Edition
EVGA SuperNOVA 750 G2 80+ GOLD 750W

Here's what else I've tested:

  • using only one stick of RAM (no difference)
  • using only one hard drive (no difference)
  • disconnecting all hard drives just to see if I could at least get a boot error on screen (no difference)

Is it possible for the PSU to be faulty, like it isn't supplying enough power to the GPU(s)? Or maybe I do actually have two broken GPUs?

Sorry this was so long, lol. Thanks for any help/advice!
 
Solution
Yeah, the temps are fine. That incident with the smoke was just the Power LED+ cable frying up, but that isn't an issue since the motherboard powers on without any of the case cables attached. I'm 100% sure now that I accidentally pulled that cable loose from its clamp, and that's why it happened.

Okay, to conclude this thread so nobody in a similar situation replies to ask if this was ever fixed: it wasn't, and it isn't going to be. I'm giving up here since it's clearly a hardware-related compatibility issue I can't pinpoint. My plan right now is to RMA the RX6800 at some point soon so I can be sure it's fine before I sell it on, and continuing to use the GT730 in the meantime until I can afford to buy an Intel/Nvidia prebuild or...

Zerk2012

Titan
Ambassador
Hey. Long time browser and I usually get most of my info on troubleshooting when I'm stuck from this website, but I've encountered a very serious issue that I need a lot of help with. Normally I can deal with these things myself, but this has me completely stumped and I'm out of ideas.

A few days back, my GPU (an RTX 2070 Super) was under relatively heavy load gaming for maybe an hour, and the computer just completely shut down in the middle of use. Since then, the PC would start, but there was no display, so I took it apart to see how it was inside when it was powered on. The GPU powers on with the fans at 100% and the usual RGB lighting, but still won't display anything.

At one point, smoke came out of the machine and I couldn't find what the cause of it, but apparently the Power LED+ cable may have been the culprit since it disconnected from it's plastic connector. This doesn't seem to have any effect powering the machine on, so I've just left it severed and disconnected the Power LED- cable from the motherboard too. I can't find any signs of anything burnt on the motherboard either (maybe I accidentally knocked the Power LED+ cable out of its socket, and it burnt up that way).

Anyway, I have a spare GPU (an RX Vega 64) and that works in very short bursts. I can sometimes get to the BIOS if I keep rebooting, but the monitor turns black again shortly after reaching it, so I can't even access it to do anything. For a little while, it got as far as the Windows login screen before dying, but it hasn't reached that since two days ago. I've tested the monitor by plugging a laptop into it, and it works fine.

I thought the PCI-E slot was the problem, so I switched the card to the other one (the x8 slot), but I still get the exact same issue with both cards (RTX2070 fans spin 100% with no output; Vega 64 only sometimes giving me very short output, and nothing otherwise). Also, I've tried HDMI and Displayport, and still have the same issue.

I don't have onboard graphics either, so I can't test that out, and I also can't test the cards out on other machines since I don't have access to any.

Here are my specs:

ASUS X470 ROG Crosshair VII Hero
Ryzen 3700x
ASUS ROG Strix GeForce RTX 2070 Super OC Edition
EVGA SuperNOVA 750 G2 80+ GOLD 750W

Here's what else I've tested:

  • using only one stick of RAM (no difference)
  • using only one hard drive (no difference)
  • disconnecting all hard drives just to see if I could at least get a boot error on screen (no difference)
Is it possible for the PSU to be faulty, like it isn't supplying enough power to the GPU(s)? Or maybe I do actually have two broken GPUs?

Sorry this was so long, lol. Thanks for any help/advice!
If you can try a friends power supply or put your card in his/her PC for testing.

One card going out OK but 2 at almost the same time makes it less likely to be a card problem.
 
Feb 14, 2022
8
0
20
If you can try a friends power supply or put your card in his/her PC for testing.

One card going out OK but 2 at almost the same time makes it less likely to be a card problem.

Yeah, I agree with you that it's unlikely both cards are the problem. The Vega64 has always been a hassle mostly due to driver problems, and would randomly restart the machine (hence why I replaced it) but it's always booted up the machine successfully regardless of how much trouble it is.

I can't try a friends power supply, but I've decided to buy another one (EVGA Supernova 750w G5) from Amazon. The other power supply I have is almost 5 years old now, so I guess it's around the time it could start failing with regular heavy use. If it doesn't work, I can just send it back and eliminate that as a possibility. I'll come back to the thread anyway after I've tried it and give an update. Thank you!
 
Feb 14, 2022
8
0
20
No luck, so it's not the PSU. I got as far as the BIOS once today after installing the new PSU, but the display output died again about two seconds in. The ASUS motherboard LED code used to give a 02 code with a white light for the VGA display section (indicating that there's a VGA issue)... but now the LED moves from 02 to something like 88 and stays there, and it kind of flickers too, so I'm not sure if it actually is 88 or not. I tried disconnecting every hard drive and moving the GPU to the other PCI-E slot again, but the outcome is still the same. I've even tried new HDMI and Displayport cables, though we can rule the monitor out as a potential issue.

Because of the new PSU coming with its own modular cables, I've replaced all of the old ones and made sure the connections are secure. I think I'll keep this one since it's smaller than the previous model, and I have

Any ideas? It seems to me it's either the CPU or the motherboard... Maybe I should get a new motherboard and CPU (with onboard graphics this time, so I don't run into this problem again). I don't really want to sink a lot of money into fixing this unnecessarily, but I guess if that's the worst scenario, it's better than buying a new GPU with how the market prices currently are. If I end up replacing the motherboard and CPU and it turns out that I just have two faulty GPUs, I'll lose it, lol.
 
Feb 14, 2022
8
0
20
I did, but they're compatible since they're EVGA G2/G5, and there was no difference in outcome anyway. Power still reaches the components since the motherboard lights up, the GPU lights are on, all of the fans spin normally, etc. (except with the RTX2070 Super, where the fans still run at full speed - the Vega 64's fans run normally, though). I researched beforehand before even attempting. Most of the EVGA modular cable types are compatible with each other, with a few exceptions.
 
Feb 14, 2022
8
0
20
Well, just as update, it's 100% the GPUs causing the issue. Both of them. I got a GT730 just to test it out, and it works fine - PC boots up with absolutely no problems. I'll RMA the 2070 Super, since that's still apparently good until sometime in 2023, so hopefully that'll be the end of this all.
 
Feb 14, 2022
8
0
20
Figured I'd come back to this. It's about 4 months on now, but I had a similar problem yesterday. I have an ASUS TUF RX6800 now since I managed to get a refund from Amazon on the 2070 Super, and was experiencing random restarts until last night when the computer went into some kind of hibernation (hibernation mode is not on btw - the monitor just had a black screen and the PC was unresponsive). The PC was then in a permanent boot loop after I shut it off, and would only sometimes get as far as the BIOS, until it wouldn't even get that far. At one point I managed to finally get it to the Windows loading screen after clearing the CMOS multiple times, but the PC kept restarting in the middle of that too, and then after a few more times of that, it wouldn't get as far as the BIOS and kept restarting over and over and over.

So I took the entire PC apart and cleaned everything. Reconnected cables etc. I even applied new thermal paste to the CPU. I've tested no RAM, one stick of RAM, etc. and tested each stick of RAM individually - no difference, still rebooting. Tested one SSD after another - again, no difference. I reinstalled the old PSU I had before - same outcome. Decided to test one last thing - removed the RX6800 and replaced it with the GT730, and everything works perfectly again. So I figured 'okay, I'll put the RX6800 into the PCIE-8X slot instead.' The PC wouldn't boot.

So I tried BOTH the GT730 in the PCIE-16X slot, and the RX6800 in the PCIE-8X slot below. The PC boots, gets as far as the login screen (slowly), and restarts every time I type in my passcode. This repeats over and over.

I've left the RX6800 in the PCIE-8X slot but disconnected it from the PSU, so the PC is back to using only the GT730. Everything is working completely fine now there's no power going to the RX6800.

Anybody any ideas what the hell is happening here? There's no way I've had three faulty GPUs giving me near identical problems. It has to be a power draw issue in some way, since the GT730 is barely taking any by comparison and working, but both PSUs are 750w and should be enough. Is there a setting in the BIOS or Windows that could cause this? Anyway, I'll just have to use the GT730 until I can figure out why the higher-demanding GPUs decide they eventually don't want to work.
 

elforeign

Distinguished
Oct 11, 2009
96
130
18,720
Figured I'd come back to this. It's about 4 months on now, but I had a similar problem yesterday. I have an ASUS TUF RX6800 now since I managed to get a refund from Amazon on the 2070 Super, and was experiencing random restarts until last night when the computer went into some kind of hibernation (hibernation mode is not on btw - the monitor just had a black screen and the PC was unresponsive). The PC was then in a permanent boot loop after I shut it off, and would only sometimes get as far as the BIOS, until it wouldn't even get that far. At one point I managed to finally get it to the Windows loading screen after clearing the CMOS multiple times, but the PC kept restarting in the middle of that too, and then after a few more times of that, it wouldn't get as far as the BIOS and kept restarting over and over and over.

So I took the entire PC apart and cleaned everything. Reconnected cables etc. I even applied new thermal paste to the CPU. I've tested no RAM, one stick of RAM, etc. and tested each stick of RAM individually - no difference, still rebooting. Tested one SSD after another - again, no difference. I reinstalled the old PSU I had before - same outcome. Decided to test one last thing - removed the RX6800 and replaced it with the GT730, and everything works perfectly again. So I figured 'okay, I'll put the RX6800 into the PCIE-8X slot instead.' The PC wouldn't boot.

So I tried BOTH the GT730 in the PCIE-16X slot, and the RX6800 in the PCIE-8X slot below. The PC boots, gets as far as the login screen (slowly), and restarts every time I type in my passcode. This repeats over and over.

I've left the RX6800 in the PCIE-8X slot but disconnected it from the PSU, so the PC is back to using only the GT730. Everything is working completely fine now there's no power going to the RX6800.

Anybody any ideas what the hell is happening here? There's no way I've had three faulty GPUs giving me near identical problems. It has to be a power draw issue in some way, since the GT730 is barely taking any by comparison and working, but both PSUs are 750w and should be enough. Is there a setting in the BIOS or Windows that could cause this? Anyway, I'll just have to use the GT730 until I can figure out why the higher-demanding GPUs decide they eventually don't want to work.

It could be the PCI-E slot on your motherboard being faulty? This was a long read, but it doesn't sound like you ever ended up changing your motherboard. If you say you saw smoke coming from the PC, my guess is something electrical happened on the PCB and burn itself out which then led to damage on the cards. If you've been using the same MB and just plugging cards in, the damage is creeping into those and why your new card failed as well.

To be safe, I would consider changing MB.
 
  • Like
Reactions: tezcatlipoca665
Feb 14, 2022
8
0
20
Apparently the Power LED+ cable was the only thing that burnt after a complete inspection, and I have a feeling it's because it wasn't securely connected (I may have knocked it out accidentally when I was changing the GPU in and out back then). If it's the PCI-E slot, it's both of them since both give me the same results, but the GT730 works fine in each (and it's PCI powered too - no direct PSU connection). I'm starting to think it's the motherboard in general some way and not just specific slots, or even potentially the combination of motherboard and CPU just not playing well with any PSU powered GPU I install and messing with power draws.

Unfortunately, I can't afford to test a new motherboard or CPU right now, so might have to just hold out with the weaker GPU until I decide to do a full upgrade combo of CPU and motherboard unless there's anything else I can try before it gets to that. If that doesn't work, I'll go insane, sell the parts, and just go with laptops from then on for the rest of my life.
 

Tac 25

Estimable
Jul 25, 2021
1,391
421
3,890
that's quite an ordeal you have with your pc.
GT 730 should at least allow you to play games at the lowest settings. Better than nothing.

may I ask, does your pc have sufficient cooling?
have you checked temps with HWmonitor during the 4 months you have your new gpu?

anyway, if something smelled burnt or smoke came out of the pc. My first instinct would be to replace the mobo and psu.
 
Last edited:
  • Like
Reactions: tezcatlipoca665
Feb 14, 2022
8
0
20
Yeah, the temps are fine. That incident with the smoke was just the Power LED+ cable frying up, but that isn't an issue since the motherboard powers on without any of the case cables attached. I'm 100% sure now that I accidentally pulled that cable loose from its clamp, and that's why it happened.

Okay, to conclude this thread so nobody in a similar situation replies to ask if this was ever fixed: it wasn't, and it isn't going to be. I'm giving up here since it's clearly a hardware-related compatibility issue I can't pinpoint. My plan right now is to RMA the RX6800 at some point soon so I can be sure it's fine before I sell it on, and continuing to use the GT730 in the meantime until I can afford to buy an Intel/Nvidia prebuild or laptop so I don't have to deal with AMD or worrying about compatibility issues for a long time.

In case this might help anybody else in a similar situation (especially with similar hardware) with their troubleshooting quest, here's a list of everything I've done:

  • using only one stick of RAM
  • testing each stick of RAM singularly
  • Memtest64 shows no errors (I've done this multiple times over the months)
  • using only one hard drive
  • disconnecting all hard drives just to see if I could at least get a boot error on screen
  • CMOS resets - sometimes several times a day
  • Default BIOS settings when I'm able to access the BIOS
  • Tested both PCI-E slots with the GT730 and RX6800
  • Tried a new EVGA SuperNOVA 750 G5PSU in place of my EVGA SuperNOVA 750 G2PSU
  • Stress tested the CPU with 3DMark to see if the CPU itself has issues
  • New thermal paste applied to CPU
  • CPU reseated
  • Disconnecting every PC case cable from motherboard and just using the motherboard's 'start' button to boot
  • Trying both the HDMI and DP connections of the RX6800 to the monitor
  • Updated motherboard BIOS just released today (8/8/2022)
  • Manually changed BIOS settings so PCI-E slot was at Gen 3 instead of Auto
  • Booted with keyboard as the only USB device
  • Uninstalled all ASUS drivers and software
  • Uninstalled AMD drivers with their own driver removal software, and with DDU software
Anyway, thanks for your help, everyone!
 
Solution