What can kill a CPU?

[Urzu1000]

I've been doing some research, and I cannot find many instances of a CPU dying of natural causes. They seem neigh indestructible under normal use.

What exactly *can* cause a CPU to die?

Why does it cause the CPU to die?

I'm just trying to fulfill some technical curiosity.

Thanks in advance!

 

[madmatt30]

Voltage, heat, static - that's about it.

All of those would be down to user error (insufficient cooling) , bad PSU's, or mains voltage spikes.

 

[Urzu1000]

But how would those actually cause the PSU to cease functioning? Would the heat melt something within the CPU? How would high voltage or static actually damage the CPU?

On some processors, people imply that running hotter (90C) can shorten the lifespan of the CPU, but I'm trying to understand why this happens.

 

[simon12]

Its to do with this https://en.wikipedia.org/wiki/Electromigration

 

[geofelt]

Processors will have thermal detection circuitry to slow down or shut off to prevent damage if temperatures get too high.

The most likely cause of failure is letting the cpu voltage get too high by overclocking.

 

[Themastererr]

Yes. The only time I have killed a CPU was when a MOSFET recently blew on my ASUS ROG motherboard. I was heavily overclocked.

When this VRM blew it sent all of the 3.3 voltage into my CPU and it was toast. I sent it to AMD for RMA.

 

[Urzu1000]

Its to do with this https://en.wikipedia.org/wiki/Electromigration

That's a very short reply, but it's exactly the information I was looking for in regards to reduced lifespan. What about static though or very high temperatures (110C+)? The Wikipedia article states that increased temperature can cause more electromigratory damage, but this would be because the cooler the metal the higher the conductivity, isn't it?

If so, having an extremely high temp should simply reduce lifespan proportionately, but I've seen from experience that it can cause huge performance issues in the short term. Static is rumored to kill components, but I don't understand why that is. Could you elaborate a little more?

Thanks and sorry for all the questions haha. I'm just trying to get a good grasp of it.

 

[Urzu1000]

Yes. The only time I have killed a CPU was when a MOSFET recently blew on my ASUS ROG motherboard. I was heavily overclocked.

When this VRM blew it sent all of the 3.3 voltage into my CPU and it was toast. I sent it to AMD for RMA.

I'll check out MOSFETs. That gives me something to go on for further research. Thanks!

 

[Themastererr]

Yes. The only time I have killed a CPU was when a MOSFET recently blew on my ASUS ROG motherboard. I was heavily overclocked.

When this VRM blew it sent all of the 3.3 voltage into my CPU and it was toast. I sent it to AMD for RMA.

I'll check out MOSFETs. That gives me something to go on for further research. Thanks!

Research everything about the VRM's or voltage regulation module's. Once learn about the different phases and quality of them you will never buy a cheap motherboard again, if you weren't already.

 

[simon12]

Its to do with this https://en.wikipedia.org/wiki/Electromigration

That's a very short reply, but it's exactly the information I was looking for in regards to reduced lifespan. What about static though or very high temperatures (110C+)? The Wikipedia article states that increased temperature can cause more electromigratory damage, but this would be because the cooler the metal the higher the conductivity, isn't it?

If so, having an extremely high temp should simply reduce lifespan proportionately, but I've seen from experience that it can cause huge performance issues in the short term. Static is rumored to kill components, but I don't understand why that is. Could you elaborate a little more?

Thanks and sorry for all the questions haha. I'm just trying to get a good grasp of it.

Sorry elaborating further is beyond my knowledge.