[SOLVED] Burnt smell after a GPU stress test, and no more video output afterwards, what to do now?

[CatastrophicBox]

Hi, first of all, i'm running a prebuilt PC (Acer Aspire GX-281) and I have very limited knowledge of hardware and PC building in general, as I haven't manipulated the inner components of a PC at all. It's not in warranty anymore, sadly. So, sorry for my novice behavior!

I have:

• GPU: NVIDIA GeForce GTX 1050
• CPU: Ryzen 7 1700
• Mobo: I don't know, it's a prebuilt :/
• RAM: 12GB

My problem started out when I did a stress test on my GPU using OCCT — after experiencing an exceptional freeze while playing civ 6, to check if the problem is related to overheating, because it has been 2 years since the PC had been cleaned, ouch, that's a huge mistake. Four minutes in, I stopped the test; two seconds after that, the PC stopped working, in a weird way: the screen was not receiving any signal and the keyboard had no lights on, the system was clearly not running anymore. However, the fans were still running and the power light was still on.
That's weird, but at least we now know that the problem is overheating! It's now time to force the computer to shut down, but... what the hell there's now a burnt smell coming from the computer?! I quickly unplugged the power cable, the burnt smell disappeared after 2 minutes.

By the way, I previously experienced many weird stuff with this computer.
First, it started while someone was playing Roblox on the PC, and suddenly, the screen goes off, no more keyboard but instead an indefinite BEEEEEEP not coming from the — turned off — speakers, but somewhere inside the computer. After restarting it, it just doesn't boot at all (no bios screen) but instead greets you with a scary and infinite BEEEEEEP until you force it to shutdown. Luckily, the warranty was still there! So, after using it, they replaced the GPU and it worked well... But the issue appeared days afterwards, so we sent it again to the customer service, they again replaced the GPU, and finally this issue did not happen anymore. Or, at least, with no more beeps...
Second, remember those weird "crashes" where the screen goes black with no OS running while the computer seems to still be running. Well, they already happened multiple times before, probably around four or five times, but it has been a while since I experienced one. It comes in two flavors: one where the computer stays on, and one where the computer just restarts. It sometimes just stops the sound or freezes it with an unbearable screeching sonority, exactly like when audio was playing while you got a sudden BSOD.

After I let the PC rest for over an hour, I decided to plug it back and see how it goes. Good news: the computer boots, and Windows fully works. Bad news: there was no video output at all, the monitor was just not receiving any signal. Although, after using the narrator to navigate through the menus (it was a pain), I saw in Task Manager that the GPU was still recognized and used, but still no output. And I'll avoid touching that computer again unless the culprit will be replaced (probably the PSU?).

So, what to do now? What could be broken, what burnt part should I look at inside the PC to see? And what to replace? Can I run the computer after cleaning it? Also, should I refrain from doing GPU stresstests in the future?

 

[TomKraven]

Right now, I'm pretty satisfied, however I'd like to upgrade my CPU later for supporting Windows 11. Also, is the death of the GPU due to the GPU itself or faulty behavior coming from another component (maybe Motherboard/PSU)? I'd like to prevent stuff like that from happening next time!

Most probably it died cause you stress tested a filthy GPU. You should clean your pc AT least every 2-3 months. If you want nice things you need to take care of them lol.

I do recommend buying a new pc all together if you have the money available. Stray away from prebuilts. Go to pcpartpicker and buy the part you need. Prioritize the PSU and GPU in you budget. CPU and Mobo need to be compatible (like a Ryzen 5600x and B550)

 

[Fatalzo]

Interesting that it was the GPU. That's extremely rare that an issue happens 4 times in a row with each time being the exact same.

 

[BoomBoomBot]

Considering the fact that 4 gpu died on you i would highly suspect that the psu is frying them. The continuous beep suggest a failure to supply power to a component

 

[CatastrophicBox]

Most probably it died cause you stress tested a filthy GPU. You should clean your pc AT least every 2-3 months. If you want nice things you need to take care of them lol.

I do recommend buying a new pc all together if you have the money available. Stray away from prebuilts. Go to pcpartpicker and buy the part you need. Prioritize the PSU and GPU in you budget. CPU and Mobo need to be compatible (like a Ryzen 5600x and B550)

Yeah you're right, that was a bad idea in the first place lol, but I thought that doing stresstests was a harmless troubleshooting tool when used moderately, but guess it isn't the case then, learned that the hard way, and in the worst time to do so.
Unfortunately I don't yet have the money to build a new PC, and to be fair I don't want to immensely screw up while building it! Though that might be an option in the future for sure, especially once GPU prices will go down.

Interesting that it was the GPU. That's extremely rare that an issue happens 4 times in a row with each time being the exact same.

Considering the fact that 4 gpu died on you i would highly suspect that the psu is frying them. The continuous beep suggest a failure to supply power to a component

About the continuous beep, it happened on startup (no bios) BUT ALSO while using the computer, screen blacks out and the beep mentally kills you! Although, that was a problem that occurred before the GPU replacements; it doesn't happen anymore.
I also rarely had— over the span of a year, and 5 or 6 times in total maybe — some sudden restarts in seemingly random moments, it could be while gaming or doing something very light! Or even no restart but instead a screen black out with no keyboard lights on, with the OS dead basically.

And thanks for your suggestions too, I should maybe try to replace the PSU with a good one (and then slowly graduate to a custom built PC?), but this is going to be my first consequent rigging with hardware in general so there's a huge chance of messing up. For now I'll just replace the broken GPU with a Gigabyte GTX 1650 I recently bought (60% more than MSRP though) which seems to be enough for my needs right now considering the prices of current GPUs, and see how it goes!

 

[BoomBoomBot]

The continuous beep with no post is the indication that a part is not receiving enough power. Usually it would occur when something isnt plugged in or plugged in incorrectly. As for a continuous beep during normal operation of the pc, im not sure what it means. The random crashes could be anything from completely harmless windows instability, software crashing or hardware issues so it is not particularly helpful info but it does point the main problem, the gpu. As for the new gpu, i wouldnt put it in the old pc. Theres a high risk that it will get damaged over time and die prematurely.

 

[TomKraven]

Yeah you're right, that was a bad idea in the first place lol, but I thought that doing stresstests was a harmless troubleshooting tool when used moderately, but guess it isn't the case then, learned that the hard way

Stress testing is like running a marathon and all the dust and filthy is like a really hot summer day. Young people will survive both easily and even grampa could run the marathon but if you put grampa(your gpu) in the marathon at 32 C he might die.