One of those sad days for a gamer

[Sassanid]

UPDATE: Problem found!


https://i.imgur.com/TxQ8ak1.jpeg
https://i.imgur.com/0LkQM2Y.jpeg
https://i.imgur.com/g9RYVrZ.jpeg

-------------------------------------------------------------------------------------------------------------------

I'm sure many of you have experienced this as avid gamers. So you've built your gaming PC, everything is working perfectly. No game crashes, everything just works as it should. Then one day comes along (for me that day is today) when out of the blue your PC starts crashing when playing your favorite game. You're not sure if it's due to some mods or some plugin, so you disable all mods and try again, still crashing, black screen, forced to hard reboot. Then it starts happening in every game. Now it happens almost instantly, start a game, crash, black screen. No way to exit except holding down the power button and rebooting.

I've spent the last hour taking apart my PC, air blowing out any dust, updated my GPU drivers to latest version, etc. But the problem persists.

It just started happening out of the blue. Do my question is, if you were me, where on earth would you start? How do I find out what's causing this when everything was working fine up until a week ago and now constant crashes. Could it be a virus? Or maybe one of my hardware components has reached the end of the road, but how to tell which one?

My specs:

10700K
zotac 3080Ti 12gb
32gb DDR4
M.2 HD
Both GPU & CPU temps always within normal range during gaming.

I appreciate any tips. What would you do if you were me?

[Shaggin'Wagon]

The 10700K was released in 2020. I don't know when you built your computer but it may have been 2020 and 4 years ago.

When is the last time you cleaned the dust out of the inside of your computer? Most people do this at least once per year but usually more often than that. If you never cleaned the dust out of your machine then that could be your problem.

[andreasaspenberg575]

it happened to me once when i played tales of arise. i just shut it down for a few weeks and when i started it back up it worked again. each part have its own fan and they collect dust. you need to wipe the dust off. fans is usually smeared wth dust so, you can not simply blow it off. you have to wipe it off. even the fan on your power supply needs cleaning.

[Sassanid]

I think I found the issue. Took a closer look at all the cables and noticed the PCI-E 8pin cables connecting my 3080ti to my PSU had the tips burned and they were giving off a very unpleasant metallic burnt smell.

The strange thing is my PC temps always seem normal and within range. CPU is watercooked and the GPU never goes above 75-77c while gaming.

I'm worried to even use this PSU again even with new cables but I also don't have the money for a new PSU rn. It's a Thermaltake 850w PSU.

Suggestions?

https://i.imgur.com/TxQ8ak1.jpeg
https://i.imgur.com/0LkQM2Y.jpeg
https://i.imgur.com/g9RYVrZ.jpeg

[Illusion of Progress]

This Black screen to restart issue is becoming concerning more frequent as time passes and pandemic era stuff is aging more. I'm seeing this on RTX 30 series and newer, and RX 6000 series and newer (especially 7800 XTs). I mean that's probably sampling bias to a point because I went through this issue, but I can't help but notice it. Literally seen it on an RTX 4060 the other day.

This happened to me on a brand new 7800 XT. Ended up pouring nearly two months of effort into it and finally did an RMA. The new one is markedly resolved of the issue, although I've found a single use case that seems to trigger it still. But it seems it's usually a hardware fault on the graphics card itself. Multiple things that can cause a PC to lose signal and restart (or require a force restart) though so there's possibly a laundry list of possible causes.

No hard data to support this but in a number of places I've tried reaching out for help or saw others with the same issue getting feedback, the possibility of bad VRAM or VRAM acting up as it passes through a warm up cycle came up. A lot of the people who had the issue had it happen once at the start of a day/session and then it's fine... until the next day. That wasn't me. I had to RMA it or else it was either fine for four days or crashing four times a day.

If that's by any chance still under warranty (?), I'd consider reaching out to support while it's got time left. Check some things first, yes, but I'd be ready to consider it.

My first recommendation to disconnect ever power related cable and the video card itself and reseat it. If you're using PCI Express risers or power cable adapters, try and remove them from the equation or replace them.

My second is that if you're overclocking, stop. This includes XMP. Test it at stock.

You can DDU and reinstall the drivers if you want. Usually drivers crash as a cascading result of faulty hardware rather than the drivers themselves, but reinstalling them will rule out a corrupt driver install itself.

Check Event Viewer. Any Event IDs 18 or 19? the former is "a fatal hardware error has occurred" and the latter is "a corrected hardware error has occurred".

Anything in the following directories?

Windows/LiveKernelReports/WHEA
Windows/LiveKernelReports/WATCHDOG

If so, WinDbg can open and analyze them.

Edit: Melted power cable! I'm panicking now, haha. Time to check mine. Yikes! Good thing you caught that.

First, I would not reuse that PSU. If it's under warranty, reach out to support. The PSU (and its cables) are not to be used anymore. If the metal contact of the power connector on the RTX 3080 is not too badly burned, it might still be fine?

[Shaggin'Wagon]

Publicado originalmente por Fret 𐎣𐎡𐎠𐎵𐎡:

Suggestions?

https://i.imgur.com/TxQ8ak1.jpeg
https://i.imgur.com/0LkQM2Y.jpeg
https://i.imgur.com/g9RYVrZ.jpeg

DO NOT USE THAT POWER SUPPLY IN A COMPUTER!

Disconnect it and replace it some how. It's now a fire hazard. Sorry but you have no choice. Some of the wires are melted on the power supply end of the cable so it's probably suffered damage inside of it that you can't see. That power supply is suspect now. Do not use.

[Sassanid]

Publicado originalmente por Illusion of Progress:

This Black screen to restart issue is becoming concerning more frequent as time passes and pandemic era stuff is aging more. I'm seeing this on RTX 30 series and newer, and RX 6000 series and newer (especially 7800 XTs). I mean that's probably sampling bias to a point because I went through this issue, but I can't help but notice it. Literally seen it on an RTX 4060 the other day.

This happened to me on a brand new 7800 XT. Ended up pouring nearly two months of effort into it and finally did an RMA. The new one is markedly resolved of the issue, although I've found a single use case that seems to trigger it still. But it seems it's usually a hardware fault on the graphics card itself. Multiple things that can cause a PC to lose signal and restart (or require a force restart) though so there's possibly a laundry list of possible causes.

No hard data to support this but in a number of places I've tried reaching out for help or saw others with the same issue getting feedback, the possibility of bad VRAM or VRAM acting up as it passes through a warm up cycle came up. A lot of the people who had the issue had it happen once at the start of a day/session and then it's fine... until the next day. That wasn't me. I had to RMA it or else it was either fine for four days or crashing four times a day.

If that's by any chance still under warranty (?), I'd consider reaching out to support while it's got time left. Check some things first, yes, but I'd be ready to consider it.

My first recommendation to disconnect ever power related cable and the video card itself and reseat it. If you're using PCI Express risers or power cable adapters, try and remove them from the equation or replace them.

My second is that if you're overclocking, stop. This includes XMP. Test it at stock.

You can DDU and reinstall the drivers if you want. Usually drivers crash as a cascading result of faulty hardware rather than the drivers themselves, but reinstalling them will rule out a corrupt driver install itself.

Check Event Viewer. Any Event IDs 18 or 19? the former is "a fatal hardware error has occurred" and the latter is "a corrected hardware error has occurred".

Anything in the following directories?

Windows/LiveKernelReports/WHEA
Windows/LiveKernelReports/WATCHDOG

If so, WinDbg can open and analyze them.

Edit: Melted power cable! I'm panicking now, haha. Time to check mine. Yikes! Good thing you caught that.

First, I would not reuse that PSU. If it's under warranty, reach out to support. The PSU (and its cables) are not to be used anymore. If the metal contact of the power connector on the RTX 3080 is not too badly burned, it might still be fine?

Yeah I'm relieved that the video card cable port seems fine, that would have been a much worse scenario.

I've got a new cable on its way to be my monday, but since the 8pin port on the PSU itself looks to be burnt too I really don't feel good about using it, who knows it might catch fire one day if this is happening!!

I'm just confused why this is even happening if all temps indicate things are ok. Cheap cables? Or some other issue.

First time this has ever happened to me in any of my builds. There's always one thing or another to keep us busy with PS gaming. It's all worth it in the end

[Sassanid]

Publicado originalmente por Shaggin'Wagon:

Publicado originalmente por Fret 𐎣𐎡𐎠𐎵𐎡:

Suggestions?

https://i.imgur.com/TxQ8ak1.jpeg
https://i.imgur.com/0LkQM2Y.jpeg
https://i.imgur.com/g9RYVrZ.jpeg

DO NOT USE THAT POWER SUPPLY IN A COMPUTER!

Disconnect it and replace it some how. It's now a fire hazard. Sorry but you have no choice. Some of the wires are melted on the power supply end of the cable so it's probably suffered damage inside of it that you can't see. That power supply is suspect now. Do not use.

You're right. I think I'll just wait and save up for a new PSU. Ahhh this sucks.

If anyone has good PSU suggestions for my rig, I'm all ears. I'll start saving now and do a little research in the mean time.

Is 850w just not enough? Maybe I should aim for a good 1000w PSU

[Sassanid]

Publicado originalmente por Alice Liddell:

Are those the original PSU cables that burned?

Yes

[emoticorpse]

It ONLY crashes when gaming? Otherwise stable?

[Sassanid]

Publicado originalmente por emoticorpse:

It ONLY crashes when gaming? Otherwise stable?

Well yes but it just started happening today and was gradually getting worse and worse and now that I opened it up and saw the burnt ends I haven't put it back together again.

Currently looking for cheap but reliable 1000w PSUs.

I don't trust using this one again after this.

[Sassanid]

Considering this Corsair. Good reviews, seems more reliable

https://www.amazon.com/Corsair-RM1000e-Modular-Low-Noise-Supply/dp/B0BYQHWJXC?th=1

[Sassanid]

Will Thermaltake be liable since it was the PSU that caused this and it is within the 5 year warranty?

Also, what if I end up finding out that the GPU has been damaged too, after I install a new PSU and test everything?

[Illusion of Progress]

Publicado originalmente por Fret 𐎣𐎡𐎠𐎵𐎡:

Yeah I'm relieved that the video card cable port seems fine, that would have been a much worse scenario.

I've got a new cable on its way to be my monday, but since the 8pin port on the PSU itself looks to be burnt too I really don't feel good about using it, who knows it might catch fire one day if this is happening!!

I'm just confused why this is even happening if all temps indicate things are ok. Cheap cables? Or some other issue.

First time this has ever happened to me in any of my builds. There's always one thing or another to keep us busy with PS gaming. It's all worth it in the end

I don't know, but please don't reuse that PSU or the cable. If the single partially singed contact of the graphics card connector can be cleaned, I'd trust it's "safe", but whether it's going to remain operational now is something you'll have to test to verify. Hopefully you don't need a new graphics card now too.

I'm not an electrician, but I know contacts heat up where there's resistance. And this heat can melt plastic connectors. That's likely what happened, and if it is, the graphics card and maybe even the PSU are "safe" and what happened was a lot of power going through it that shouldn't have (can you clarify if you were pigtailing this?), but you definitely need a new cable and I'd just get a new PSU since you don't want to chance cables that aren't verified to work with a given PSU, and the PSU connector is pretty rough anyway.

This is what's doing the RTX 4090 cables in but I don't think I've seen it on too many other GPUs, but I have seen it a few times. They just keep pushing the size and power draw of these things and this is the result. Then they blame user error. In reality, imperfect contacts or cable bends have probably always been happening, but graphics cards weren't using as much power on average so it's only now becoming an issue. That's my guess.

Make sure the new cables have firm contact and that there is slack with the cables. Don't pull them tight or do any sharp bends right after where it connects to the graphics card.

The concerning part is why the cable burned on both ends. I've seen that posted on Reddit before though. It's not just the new 12VHPWR connectors doing this (they're just the ones doing it the most).

[Electric Cupcake]

When you've eliminated the impossible, what remains...

[Sassanid]

Publicado originalmente por Illusion of Progress:

Publicado originalmente por Fret 𐎣𐎡𐎠𐎵𐎡:

Yeah I'm relieved that the video card cable port seems fine, that would have been a much worse scenario.

I've got a new cable on its way to be my monday, but since the 8pin port on the PSU itself looks to be burnt too I really don't feel good about using it, who knows it might catch fire one day if this is happening!!

I'm just confused why this is even happening if all temps indicate things are ok. Cheap cables? Or some other issue.

First time this has ever happened to me in any of my builds. There's always one thing or another to keep us busy with PS gaming. It's all worth it in the end

I don't know, but please don't reuse that PSU or the cable. If the single partially singed contact of the graphics card connector can be cleaned, I'd trust it's "safe", but whether it's going to remain operational now is something you'll have to test to verify. Hopefully you don't need a new graphics card now too.

I'm not an electrician, but I know contacts heat up where there's resistance. And this heat can melt plastic connectors. That's likely what happened, and if it is, the graphics card and maybe even the PSU are "safe" and what happened was a lot of power going through it that shouldn't have (can you clarify if you were pigtailing this?), but you definitely need a new cable and I'd just get a new PSU since you don't want to chance cables that aren't verified to work with a given PSU, and the PSU connector is pretty rough anyway.

This is what's doing the RTX 4090 cables in but I don't think I've seen it on too many other GPUs, but I have seen it a few times. They just keep pushing the size and power draw of these things and this is the result. Then they blame user error. In reality, imperfect contacts or cable bends have probably always been happening, but graphics cards weren't using as much power on average so it's only now becoming an issue. That's my guess.

Make sure the new cables have firm contact and that there is slack with the cables. Don't pull them tight or do any sharp bends right after where it connects to the graphics card.

The concerning part is why the cable burned on both ends. I've seen that posted on Reddit before though. It's not just the new 12VHPWR connectors doing this (they're just the ones doing it the most).

It's scary because this could have easily resulted in a fire and much worse and there was no indication that there was anything wrong or temps too high, everything was fine and then suddenly this happens.

I'm going to contact thermaltake about this and see what they say, since the power supply is still within the 5 year warranty