[SOLVED] Questioning if PSU still safe

Craftsman_2222

Honorable
Nov 4, 2015
48
0
10,560
So I'm running a decently old set up (i7-5820k, 980ti, Intel PCIe SSD/ Sata Samsung Evo) and my GPU just decided to burn itself out when playing God of War (which anyone know how to fix the crashing as well). The game would randomly crash and restart my PC. When the card burnt out the PC went into a small boot loop which is when I started to smell it.

Im looking at getting a new card (6600xt) but I'm worried that the issue isnt resolved and could potentially happen again. As far as I'm concerned my PSU is the only part which could have an issue in the future as it seems like the card failed cause of power surge. Should I think about getting a new PSU as well? How safe is it to continue with the same one. Its an HX850i.
 
Solution
There's multiple affects stemming from a single issue. That HXi has plenty of good protections built in, so I seriously doubt anything the gpu did is affecting other hardware. Your bootloop is most likely stemming from data loss due to sudden shutdown, not related to psu ability.

Psus don't supply max power or supply max voltages or other such. All they supply is demanded power. So if your whole pc only used 400w, wouldn't matter if it was powered by a 450w psu or a 1600w psu, the pc only demands 400w. So that's what it gets. If the gpu fried because of power surge, it fried because it demanded the psu supply the power, not because the psu suddenly decided to overload the gpu.

JeffreyP55

Distinguished
Mar 3, 2015
572
142
19,070
So I'm running a decently old set up (i7-5820k, 980ti, Intel PCIe SSD/ Sata Samsung Evo) and my GPU just decided to burn itself out when playing God of War (which anyone know how to fix the crashing as well). The game would randomly crash and restart my PC. When the card burnt out the PC went into a small boot loop which is when I started to smell it.

Im looking at getting a new card (6600xt) but I'm worried that the issue isnt resolved and could potentially happen again. As far as I'm concerned my PSU is the only part which could have an issue in the future as it seems like the card failed cause of power surge. Should I think about getting a new PSU as well? How safe is it to continue with the same one. Its an HX850i.
You have an old setup. The power supply is the heart of your system. If it has a heart attack it could take out the whole system. Find out were the ozone smell is coming from. I bought the best 1200w supply I could purchase a few years back. With the power hunger of newer GPU it is a good thing I did. The supply has a 1175w 12v rail. There is a sequence of events for trouble shooting.
 

Craftsman_2222

Honorable
Nov 4, 2015
48
0
10,560
You have an old setup. The power supply is the heart of your system. If it has a heart attack it could take out the whole system. Find out were the ozone smell is coming from. I bought the best 1200w supply I could purchase a few years back. With the power hunger of newer GPU it is a good thing I did. The supply has a 1175w 12v rail. There is a sequence of events for trouble shooting.
This is why I want to ensure the power supply is still good. Its a quality power supply and I really need to save money right now. I know its the card that failed as I can see the DrMOS chip that failed. How do I test if its still good?
 

JeffreyP55

Distinguished
Mar 3, 2015
572
142
19,070
This is why I want to ensure the power supply is still good. Its a quality power supply and I really need to save money right now. I know its the card that failed as I can see the DrMOS chip that failed. How do I test if its still good?
You need a test machine or expensive testing equipment. As layman all we can do is shotgun a problem. Do you have a buddy that could possibly help you? If you can afford it, there is nothing wrong with collecting old parts for future test purposes. Your setup could get away with a fairly inexpensive 550 watter. Bought one for my wife's business to keep her old 2 core Pentium working...
When I was designing PPS equipment for SLAC (Stanford Linear Accelerator) I doubled up on power and thermal requirements.
 

Karadjgne

Titan
Ambassador
There's multiple affects stemming from a single issue. That HXi has plenty of good protections built in, so I seriously doubt anything the gpu did is affecting other hardware. Your bootloop is most likely stemming from data loss due to sudden shutdown, not related to psu ability.

Psus don't supply max power or supply max voltages or other such. All they supply is demanded power. So if your whole pc only used 400w, wouldn't matter if it was powered by a 450w psu or a 1600w psu, the pc only demands 400w. So that's what it gets. If the gpu fried because of power surge, it fried because it demanded the psu supply the power, not because the psu suddenly decided to overload the gpu.
 
Solution

JeffreyP55

Distinguished
Mar 3, 2015
572
142
19,070
There's multiple affects stemming from a single issue. That HXi has plenty of good protections built in, so I seriously doubt anything the gpu did is affecting other hardware. Your bootloop is most likely stemming from data loss due to sudden shutdown, not related to psu ability.

Psus don't supply max power or supply max voltages or other such. All they supply is demanded power. So if your whole pc only used 400w, wouldn't matter if it was powered by a 450w psu or a 1600w psu, the pc only demands 400w. So that's what it gets. If the gpu fried because of power surge, it fried because it demanded the psu supply the power, not because the psu suddenly decided to overload the gpu.
It doesn't matter. The same amount of current will be used whatever the power rating the power supply is. I was trying to give the man a reference point. So, what has he done to rectify the issue?
Where does the ozone smell emanate from? The power supply, motherboard or GPU? He is not doing the basics it takes to trouble shoot a horses petute.
A decently designed PS with quality parts should last for years and years. Discrete components could last a lifetime. Dried electrolytic cheap caps could be the issue.
 
Last edited: