glocked89
Had my card since launch. For older titles I've been undervolting, 1800mhz 0.850v. I'll only load up my oc profile for cp2077.

Could you not just set a global fps limit to achieve less draw/ load on the card?

timmyboy04

glocked89
Had my card since launch. For older titles I've been undervolting, 1800mhz 0.850v. I'll only load up my oc profile for cp2077.

Could you not just set a global fps limit to achieve less draw/ load on the card?

I have a fps in place as well. With just the fps cap, the card will still pull more than 0.850v at times. With the undervolt it'll never draw more than 0.850v.

I am also undervolting my card even if I haven't experienced any kind of graphic crashes due to over voltage (more due to unstable OC when overclocking)...
 
For normal, daily gaming use, undervolting basically helps keep the same performance or even greater while keeping voltages and temperature way lower. Currently using a custom curve at max 0.950 V with core boost clocks between 1980 and 1965MHz and temps stable at max 51º C (Hybrid cooling) while playing Cyberpunk 2077.
 
It also helps controlling the PCI slot power draw "issue" since it rarely reaches 70 W.

timmyboy04

glocked89
Had my card since launch. For older titles I've been undervolting, 1800mhz 0.850v. I'll only load up my oc profile for cp2077.

Could you not just set a global fps limit to achieve less draw/ load on the card?

The problem isn't the power draw, the problem is it's spiking voltages up to max-rating at clocks and usage numbers it shouldn't be at.  Likely, this means certain components that aren't supposed to "see" this sort of voltage are rapidly degrading on the card, because too much voltage is being shoved through them unexpectedly.

After another day of gaming I'm 100% confident this is the issues with these cards.

To mitigate the issue, I've set the card to run at STOCK clocks above 1.062V, and boost below that.  That means I still get OC available clocks as long as the card is cool enough, without worrying about having to change profiles for different games.

A firmware update should be able to resolve this.

Rewire92

timmyboy04

glocked89
Had my card since launch. For older titles I've been undervolting, 1800mhz 0.850v. I'll only load up my oc profile for cp2077.

Could you not just set a global fps limit to achieve less draw/ load on the card?

The problem isn't the power draw, the problem is it's spiking voltages up to max-rating at clocks and usage numbers it shouldn't be at.  Likely, this means certain components that aren't supposed to "see" this sort of voltage are rapidly degrading on the card, because too much voltage is being shoved through them unexpectedly.

After another day of gaming I'm 100% confident this is the issues with these cards.

To mitigate the issue, I've set the card to run at STOCK clocks above 1.062V, and boost below that.  That means I still get OC available clocks as long as the card is cool enough, without worrying about having to change profiles for different games.

A firmware update should be able to resolve this.

Can you elaborate on this I don't really understand what you mean. Do you have a photo of the voltage curve?

mricypaw1

Rewire92

timmyboy04

glocked89
Had my card since launch. For older titles I've been undervolting, 1800mhz 0.850v. I'll only load up my oc profile for cp2077.

Could you not just set a global fps limit to achieve less draw/ load on the card?

The problem isn't the power draw, the problem is it's spiking voltages up to max-rating at clocks and usage numbers it shouldn't be at.  Likely, this means certain components that aren't supposed to "see" this sort of voltage are rapidly degrading on the card, because too much voltage is being shoved through them unexpectedly.

After another day of gaming I'm 100% confident this is the issues with these cards.

To mitigate the issue, I've set the card to run at STOCK clocks above 1.062V, and boost below that.  That means I still get OC available clocks as long as the card is cool enough, without worrying about having to change profiles for different games.

A firmware update should be able to resolve this.

Can you elaborate on this I don't really understand what you mean. Do you have a photo of the voltage curve?

He posted the photos in the original post. 

for what its worth, im looking over the logs of my 3090 ftw hybrid right before it died, from back in december.  
 
while playing witcher 3, it spiked up to 1.075v for a couple of seconds, and then instantly dropped back to .0738v.  the power consumption seemed to be all over the place too.   while at 1.075v it was drawing between 189-269w. and then 380w @ 1.056v
 
the last entry before she blew up is 397.989w @ 1.038v

Rewire92

A firmware update should be able to resolve this.

Wow thanks for figuring all this out.  If they could fix with a firmware update, would that be a case of limiting the card's performance to achieve safety?

Good job! If it's a design flaw, then ALL 3080s and 3090s are a ticking time bomb and they will only fix this with a new revision. Probably a new bios should "fix" this at the expense of the bleeding edge performance and oc. Sad thing... And they are waiting way too long to release at least a beta bios that addresses the problem, at least to stop killing more boards.
Anyone tried running those games without any oc software to see if the problems continue? JayzTwoCents had overclocking problems using Precision X1 with 3090 FTW3, which were solved using Afterburner. So maybe overclocking software could trigger those spikes by causing some sort of interference/conflict? (sorry if I'm talking nonsense, I'm not an expert by any means)

I'm hopeful EVGA will release a bios update soon which will prevent the voltage spikes. Hopefully it can be solved through that rather than a hardware revision.