4060 Ti graphics card catches fire and melts like an ice cream — user claims they were using it for simple Adobe work and PUBG

(Image credit: user_634494 / Quasar Zone)

A South Korean gamer says that their two-year-old RTX 4060 Ti reportedly caught fire and melted while being used for gaming. According to their post on Quasar Zone [machine translated], the GPU wasn’t overclocked, and they weren’t using it for anything heavy. They primarily used their computer for simple tasks on Adobe, and the most demanding game they played was PUBG. Although high-powered Nvidia GPUs like the RTX 5090 and 5080 with their 12VHPWR and 12V-2x6 power connectors were known for melting, lower-powered graphics cards like the RTX 4060 Ti are not prone to this issue.

“I’d been using it for about two years, and suddenly, while gaming, the computer shut off. I thought, “What’s going on?” and pressed the power button to turn it back on. Smoke started coming out and then flames burst out — I quickly unplugged everything, turned off the power, and disassembled the PC,” said the anonymous user. They also added, “About three months ago, I was in a rush for work and called a repair shop, expecting to get ripped off. They replaced the motherboard, and the computer started to work again. Since I was in a hurry, I accepted the refurbished unit they offered — now I’m wondering if this was a motherboard issue.”

Apparently, the user had an issue with their computer and had their motherboard replaced with a refurbished one a few months before this incident. They now suspect that the GPU bursting into flames was caused by the replacement, although it’s unlikely to happen. The damage to the motherboard shows burn marks on the backside, near one of the screws that secures the heat sink on the GPU. There’s also a massive melt pattern near where the LR22 inductors used for the GPU’s voltage regulation are placed. You can see their location in 4060 Ti teardown videos, and the damage seen in the photographs indicates that the fire was likely caused by a VRM failure.

Tear-Down Speedrun of NVIDIA RTX 4060 Ti Founders Edition - YouTube

Watch On

The user said that they’d sent their computer back to the repair shop to be looked at, especially since the motherboard was still in warranty. However, given that the damage was on the GPU, they’re worried that it might not be covered by that.

This is such an unfortunate and potentially dangerous incident, and one that we did not expect, given that the GPU wasn’t even a high-powered one. One commenter suggested that the screw with the most damage might have been over-torqued during manufacture, resulting in a short that led to the failure of the VRMs. However, this is pretty unlikely, especially since the GPU has been in use for a couple of years without any issues. Although the GPU did not come from an international brand like Asus or Gigabyte, comments said that Emtek was a decent South Korean GPU brand.

We hope that this is an isolated case and that the GPU manufacturer addresses the issue quickly. After all, with many gamers rocking this GPU in their systems, it’s going to be a disaster if these start going up in flames during regular use.

Follow Tom's Hardware on Google News to get our up-to-date news, analysis, and reviews in your feeds. Make sure to click the Follow button.

See more GPUs News

Jowi Morales is a tech enthusiast with years of experience working in the industry. He’s been writing with several tech publications since 2021, where he’s been interested in tech hardware and consumer electronics.

12 Comments Comment from the forums

bit_user

What's the role of underwriters, in an incident like this? If an organization like UL certifies a product and it catches fire, are they on the hook in any way? Like if the fire is shown to be due to a design flaw that they should've caught?

Or, is the entire value in such a certification simply the sense of security it's meant to give consumers?
Reply
King_V

Forbidden ice-cream.
Reply
wakuwaku

One commenter suggested that the screw with the most damage might have been over-torqued during manufacture, resulting in a short that led to the failure of the VRMs. However, this is pretty unlikely, especially since the GPU has been in use for a couple of years without any issues.
If you had use normal Google Translate, instead of Tom's AI to translate, you would know that the commenter mentioned that possible over-torquing of the screw might have led to the board WARPING, in turn causing a short. Warping of a pcb doesn't immediately lead to a short, heck it may not happen at all depending on tolerances. But the warping can can worsen overtime especially in an environment where the board is hot then cool then hot again in short bursts to medium length bursts repeatedly over the years, aka using a GPU normally.
See how a different understanding using PROPER tools changes something unlikely to something more likely.
Reply
Eximo

bit_user said:
What's the role of underwriters, in an incident like this? If an organization like UL certifies a product and it catches fire, are they on the hook in any way? Like if the fire is shown to be due to a design flaw that they should've caught?

Or, is the entire value in such a certification simply the sense of security it's meant to give consumers?
If anything it makes companies less liable. They are basically saying that rigorous testing was performed and that any such occurrence could be deflected from their product. Maybe a faulty component or something else. At least that would be what the lawyers would argue.
Reply
bit_user

Eximo said:
If anything it makes companies less liable.
Is this a hypothetical arguement, or based on actual and specific knowledge/experience?

My company has had to subject our products for various testing (RF, I'm not sure exactly what else) and they don't always pass on the first try.
Reply
Eximo

bit_user said:
Is this a hypothetical arguement, or based on actual and specific knowledge/experience?

My company has had to subject our products for various testing (RF, I'm not sure exactly what else) and they don't always pass on the first try.
On UL's page as one of the benefits of participating. Generally also how lawyers defend clients, place the blame somewhere else, if possible.

Not sure what personal pass/fail of UL certification would have to do with it. On a more general note, failing would be non-UL compliant making a company certainly more liable. Passing would say that, to the best of their knowledge, the product is safe and could then look for other causes.
Reply
bit_user

Eximo said:
On a more general note, failing would be non-UL compliant making a company certainly more liable.
If you fail, then you make design changes and re-submit. If a company deems these certifications to be worthwhile (and indeed, where they're required by law), you don't ship the product until it passes.

Eximo said:
Maybe a faulty component or something else. At least that would be what the lawyers would argue.
UL certification certainly takes into account the possibility of component failure. If a component failure can create a fire or other hazard, it's an unsafe design almost by definition, because components can and do fail.
Reply
Eximo

bit_user said:
If you fail, then you make design changes and re-submit. If a company deems these certifications to be worthwhile (and indeed, where they're required by law), you don't ship the product until it passes.

UL certification certainly takes into account the possibility of component failure. If a component failure can create a fire or other hazard, it's an unsafe design almost by definition, because components can and do fail.

Yes, I am understanding that. I'm not sure what relevance it has here. Either this card passed certification or it didn't. If it did, hooray! If it didn't, or this particular version of the card didn't, it would make the company more liable. Which was your question.

Yes, again, we are talking about liability. A smart lawyer sues everyone. UL, the manufacturer, the manufacturer of sub components, etc. The defense for UL would say they did their due diligence with the samples provided for testing. The manufacturer would say they did due diligence in getting that certification. The sub component manufacturer would try to blame the manufacturer. Maybe even talk about power surges or a faulty power supply. It is up to the plaintiff here to assign the blame using facts, and to prove damages.

In this case, if the PC is damaged and is not replaced by the GPU manufacturer, in addition to the faulty card, they could go after additional damages. But generally this is not cost feasible for small sums. What with arbitration clauses and potential small claims, probably easier to just go through home/renter's insurance.
Reply
bit_user

Eximo said:
Yes, again, we are talking about liability. A smart lawyer sues everyone. UL, the manufacturer, the manufacturer of sub components, etc.
It's not what the plaintiff does that's in question. My question was about the value of a UL certification, which could either include liability insurance or be required for liability insurance.

If you don't know, that's okay. This is not school, where it's usually worth trying to BS your way through an answer. Opting to say nothing avoids putting wrong information out into the world and therefore has value when there's a good chance of being wrong.
Reply
Eximo

bit_user said:
It's not what the plaintiff does that's in question. My question was about the value of a UL certification, which could either include liability insurance or be required for liability insurance.

If you don't know, that's okay. This is not school, where it's usually worth trying to BS your way through an answer. Opting to say nothing avoids putting wrong information out into the world and therefore has value when there's a good chance of being wrong.
I think you are confused as to which of your four questions I am attempting to answer. The first paragraph's question is what I was considering, liability.

I think I agree with you that the value to the consumer is just a warm fuzzy feeling. In practice the only thing that is going to change anything would be class action, individual protections are limited.

Historically, I can imagine that before UL certification there was some pretty terrible stuff out there. And these days, a lot of UL certifications are faked for simply copied from existing devices to appear legitimate.
Reply

Show more comments