check your PSU

Just an update.
I have again started to get random shutdowns. But not full shutdowns. LEDS stay on.
I think the replacement 750W was also botched or had bad quality.
I'm going to send it back for another one.
 
Kinda annoyed that Seasonic claims they "ran out" of higher power wattages when I offered to pay the difference to get a higher rated as replacement.
 

tamalero
Just an update.
I have again started to get random shutdowns. But not full shutdowns. LEDS stay on.
I think the replacement 750W was also botched or had bad quality.
I'm going to send it back for another one.
 
Kinda annoyed that Seasonic claims they "ran out" of higher power wattages when I offered to pay the difference to get a higher rated as replacement.
 

750w isnt even close to being enough.
Power Spikes that dont Show up in messurements are way higher.
Get a 1000w PSU, that will fix your Problem 100%

Itssladenlol
 
750w isnt even close to being enough.
Power Spikes that dont Show up in messurements are way higher.
Get a 1000w PSU, that will fix your Problem 100%

Was more than 2 months with no issue.
And, these Seasonic are supposed to handle over 750W transitions. Hence why they were replacing the old ones that could not handle the spikes with the newer version.
 

N

jankerson
 
 
It's not so much the wattage than it is the exact models of PSUs. ;)
 
Could be 750W as long as it's one of the models that is known to work it's fine.
 
It's just certain models handle the trainset power spikes better than others.
 

True...  Could also be a 650.. as I ran my 3090 FTW3 Ultra and i9-9900k on one and it was daisy chained.

Just spent 2 weeks dealing with this. 
 
Before the 3080FTW3, rock solid system, rock solid PSU. 850watt toughpower grand 850watt, driving a 1080ti FTW3 with an i7 7800x, no overclocking, ever. 
After 3080FTW3, PC would power off/power on during games, anywhere from 5 mins to 2 hours into gameplay. Rebuilt system during troubleshooting.
Reducing GPU power to 90%, 80%, 70% would slow down the reboots, but they would still occur even at 70% power. Voltage monitoring of system showed nothing that would cause this, at least at a logging level. (variance on main 12v rail was less than .1v, infrequently)
 
EVGA offered an RMA, but I had already put a new 1000watt PSU in shipping. (EVGA supernova G5, chinese import). The new PSU stopped the rebooting entirely, and the system has now run for 3+ days without a single problem. 
The toughpower grand 850watt is still a good PSU, tests good, and runs every other configuration just fine. 

Whose issue was this? Is this a defective GPU that can be stabilized with a large extra current overhead? Or was this a defective PSU that could only be determined under THIS ONE USE CASE? It's anyone's guess.
 

milkman76
Just spent 2 weeks dealing with this. 
 
Before the 3080FTW3, rock solid system, rock solid PSU. 850watt toughpower grand 850watt, driving a 1080ti FTW3 with an i7 7800x, no overclocking, ever. 
After 3080FTW3, PC would power off/power on during games, anywhere from 5 mins to 2 hours into gameplay. Rebuilt system during troubleshooting.
Reducing GPU power to 90%, 80%, 70% would slow down the reboots, but they would still occur even at 70% power. Voltage monitoring of system showed nothing that would cause this, at least at a logging level. (variance on main 12v rail was less than .1v, infrequently)
 
EVGA offered an RMA, but I had already put a new 1000watt PSU in shipping. (EVGA supernova G5, chinese import). The new PSU stopped the rebooting entirely, and the system has now run for 3+ days without a single problem. 
The toughpower grand 850watt is still a good PSU, tests good, and runs every other configuration just fine. 

Whose issue was this? Is this a defective GPU that can be stabilized with a large extra current overhead? Or was this a defective PSU that could only be determined under THIS ONE USE CASE? It's anyone's guess.
 

 You could say the issue is that modern GPUs are being built and configured to be more and more aggressive in their boosts to drive higher performance.
They pull way too much power now in an instant, triggering the protection of older PSUs that did not expect such an instant demand.
 
Anyway, in my case...  been trying to a decent find 1000W PSUs in my country. And the only ones available are crappy thermaltake and older also affected coolermaster units x_x
 

tamalero

milkman76
Just spent 2 weeks dealing with this. 
 
Before the 3080FTW3, rock solid system, rock solid PSU. 850watt toughpower grand 850watt, driving a 1080ti FTW3 with an i7 7800x, no overclocking, ever. 
After 3080FTW3, PC would power off/power on during games, anywhere from 5 mins to 2 hours into gameplay. Rebuilt system during troubleshooting.
Reducing GPU power to 90%, 80%, 70% would slow down the reboots, but they would still occur even at 70% power. Voltage monitoring of system showed nothing that would cause this, at least at a logging level. (variance on main 12v rail was less than .1v, infrequently)
 
EVGA offered an RMA, but I had already put a new 1000watt PSU in shipping. (EVGA supernova G5, chinese import). The new PSU stopped the rebooting entirely, and the system has now run for 3+ days without a single problem. 
The toughpower grand 850watt is still a good PSU, tests good, and runs every other configuration just fine. 

Whose issue was this? Is this a defective GPU that can be stabilized with a large extra current overhead? Or was this a defective PSU that could only be determined under THIS ONE USE CASE? It's anyone's guess.
 

You could say the issue is that modern GPUs are being built and configured to be more and more aggressive in their boosts to drive higher performance.
They pull way too much power now in an instant, triggering the protection of older PSUs that did not expect such an instant demand.
 
Anyway, in my case...  been trying to a decent find 1000W PSUs in my country. And the only ones available are crappy thermaltake and older also affected coolermaster units x_x
 

But what is that saying, still? Whose fault is it? I've been building for 30 years, and it is exceedingly rare I run into any tech issue I cannot track back to a vendor.

Unless I have a bench with power monitoring systems that I currently do not have access to, I cannot prove when the "aggressive power draw" occurred on the GPU, or when the PSU was unable to handle the transition to a higher draw. Existing runtime monitoring software and BIOS level monitoring do not show anything that would cause a reboot or failure, so the issue isnt something a customer... can easily prove. This stinks, and one of the vendors should step up and try a little honesty for 2 seconds, if that's possible under an economic system that makes everyone lie to cover their butts. 
 
I am only half serious when I say Im not sure whose issue it is. If I were a betting man, I wouldnt bet that nearly EVERY PSU maker on the planet has the same "faulty" PSUs, that can only be shown to be faulty under this one use case. (EVGA 3080 and 3090s). I dont see any FEs with this issue. 
 
EDIT: Oh, and good luck with your PSU!!! Thermaltake tried to fight with me, at first, and thought they were talking to someone who would give up with "if the PSU turns on, it isnt bad." They tried to refuse my 10 year warranty on that PSU until I threatened them. Now, I have an open RMA with thermaltake for that grand 850watt. 
 

milkman76
 
But what is that saying, still? Whose fault is it? I've been building for 30 years, and it is exceedingly rare I run into any tech issue I cannot track back to a vendor.

Unless I have a bench with power monitoring systems that I currently do not have access to, I cannot prove when the "aggressive power draw" occurred on the GPU, or when the PSU was unable to handle the transition to a higher draw. Existing runtime monitoring software and BIOS level monitoring do not show anything that would cause a reboot or failure, so the issue isnt something a customer... can easily prove. This stinks, and one of the vendors should step up and try a little honesty for 2 seconds, if that's possible under an economic system that makes everyone lie to cover their butts. 
 
I am only half serious when I say Im not sure whose issue it is. If I were a betting man, I wouldnt bet that nearly EVERY PSU maker on the planet has the same "faulty" PSUs, that can only be shown to be faulty under this one use case. (EVGA 3080 and 3090s). I dont see any FEs with this issue. 
 
EDIT: Oh, and good luck with your PSU!!! Thermaltake tried to fight with me, at first, and thought they were talking to someone who would give up with "if the PSU turns on, it isnt bad." They tried to refuse my 10 year warranty on that PSU until I threatened them. Now, I have an open RMA with thermaltake for that grand 850watt. 
 

Its incorrect to say that "only" EVGA cards were to blame here. I have been on reddit and other forums and ANY overclocked 3080 and 3090 seem to have some sort of issues with very specific older power supplies.
Most of these PSUs already been replaced by newer more sturdy ones.
And I still remember back then when the 1080's and TI's caused similar issues with even older PSUs.
Let's not forget how Nvidia had to make a driver to make sure the 3000's series weren't pushing too aggressively on their boosts (which caused many cards to bsod, freeze or reboot in he first weeks of release).
 
In the other hand I just installed the replacement 750W from Seasonic.
What is interesting is, despite being the same model, brand, type and code. The sticker is different and has different position for a few things. Like the 80 Titanium plus. I wonder if I just got the latest version of the same model.
 
Anyway, If I'm getting the same issue. I'm going to sell the PSU and just get a bigger more powerful one.
 
 

I had to rma my older focus model, seasonic sent mr an updated one and it works fine (850 gx)