Dead EVGA 3090 FTW3 Ultra Gaming?

Canon

2[H]4U
Joined
Aug 12, 2004
Messages
2,561
I currently have a support ticket open with EVGA, I'm assuming RMA at this point since I can't get any vide signal at all out of the card now, but I wanted to see if anyone at [H] has has similar experiences with this card. There seems to be a serious sprinkling of this issue reported on the EVGA community forums and Reddit from what I have been able to find in my search for an answer.

Keep in mind I am not doing any overclocking, in fact I haven't even bothered since I got this card at the end of November.

I came across this thread over on the EVGA forums which really got me thinking about this, because my crashing was only really observed with less demanding games (GTA V, Valheim for me). I literally played CyberPunk 2077 for > 100 hours with no issues whatsoever. Really bad if Rewire92's theory is correct regarding a possible issue with low load causing high voltage spikes, possibly killing the card.
https://forums.evga.com/Fixing-EVGAs-7-Figure-Problem-with-FTW3-30-Series-cards-m3217284.aspx

Anyhow, here's what I observed on my system and what I did to attempt troubleshooting if anyone is interested.

1. Playing a game, screen goes black and my monitor reports no Display Port signal, GPU fans ramp up to 100%. System seems to still be responding (sound still working, keyboard responds, OS shuts down normally with the power button. The GPU LEDs look normal (ie: my customized LED effects are visible, no red LEDs by the PCI-E connectors). When this happens I observed in the windows reliability history a LiveKernelEvent code 141 error is logged. Here is a copy and paste from that error:

Source
Windows

Summary
Hardware error

Date
‎2/‎22/‎2021 7:43 PM

Status
Not reported

Description
A problem with your hardware caused Windows to stop working correctly.

Problem signature
Problem Event Name: LiveKernelEvent
Code: 141
Parameter 1: ffffc00cd1dc5050
Parameter 2: fffff801491e71c0
Parameter 3: 0
Parameter 4: 1b84
OS version: 10_0_19042
Service Pack: 0_0
Product: 256_1
OS Version: 10.0.19042.2.0.0.256.48
Locale ID: 1033

2. Playing a game, GPU fans go to 100%, game continues to run. Checking temperatures in EVGA Precision X1 reveals normal GPU temperatures and then the temperatures go down rapidly since the fans are running at 100%. After a few minutes, this seems to clear and fans return to normal operation. When I checked EVGA Pecision X1 when this was happening, I found that the main screen showed fan #3 at 98% with Fan#1 and Fan#2 at 0% (which was wrong, I could see them spinning), going to the hardware tab revealed Fan#1,Fan#2 at 51% and Fan#3 at 98%.

3. Most recently I had no active 3D applications running and was just using my web browser when the fans on the GPU ramped to 100% and the screen went black with no Display Port signal. Ever since that time I am unable to get any display output from the GPU on any of the DP connections, including BIOS post screens which I normally see after rebooting from the previous issues listed above. GPU looks normal, custom LED effects are visible, no RED LEDs illuminated by the PCI Express connectors, fans are spinning.

Over time, things I have tried to troubleshoot these issues:

1. Uninstall Nvidia Drivers/Geforce Experience, run DDU in safe mode with network cable disconnected, reload current Nvidia drivers. Most recently 461.40.

2. Step 1 with multiple different older drivers 461.33, 460.79, 460.89, 461.09.

3. Uninstall EVGA Precision X1.

4. Ensure no hardware monitoring utilities are running, issue still happened.

5. Power off PC, toggle between the OC and Quiet Bios modes, start PC, restart PC for the new BIOS settings to apply. I have had the issues with both BIOS switch settings.

6. Remove and re-seat the GPU and power connectors.

7. Tried another known working Corsair RM 750 PSU, same issue with both PSUs (my current PSU is a Corsair HX 850).

Airflow in my Fractal Design Mesifhy 2 case is ample with 3x 140MM Noctua fans and 140mm Exhaust. GPU temps on the OC bios were 67-68 degrees C while gaming, 74-76 degrees C on the quiet mode.

I have swapped in my Nvidia Geforce RTX 2080 Ti FE card and it is working perfectly fine with no issues. The EVGA GeForce RTX 3090 FTW3 Ultra Gaming GPU seems to be dead at this point.
 

polonyc2

Fully [H]
Joined
Oct 25, 2004
Messages
19,565
I'm hearing a lot of red light of death reports from 3090 FTW3 Ultra owners...I have a 3080 FTW3 Ultra but haven't installed it yet...
 

Canon

2[H]4U
Joined
Aug 12, 2004
Messages
2,561
No other PSUs to try. The HX 850 is literally brand new, the RM 750 is about 5 years. It happily ran my system with an RTX 2080ti and this card as well while I was playing CyberPunk 2077. I was going to buy a higher wattage PSU, but that is literally all I have been able to find in stock. I settled on a decently sized PSU of high quality (the HX850 is 80 plus platinum). It is a high quality PSU and the system draws ~560-600 watts under load from my testing. I am not hitting OCP or anything, like I said the system is still powered on and working after the GPU goes haywire. If I was hitting OCP, I would expect a hard power off.

Also, for those wondering why I have an RTX 3090 in a 5 year old i7 6700k system... still waiting on stock for a new AMD R9 processor... I have all the parts, but no CPU to complete the build...

Like I said above, I literally played CyberPunk for well over 100 hours without an issue, if PSU was an issue, that would have done it... That game pretty much maxed out my CPU and GPU most of the time while playing.

EVGA has asked me to begin the RMA process. Guess we'll see how this goes.

I mostly opened this thread to see if others are having similar issues with this card here at [H]. I don't think any amount of hardware swapping is going to make this GPU work considering the rest of my components are working fine and this same card was working before.
 

pendragon1

Fully [H]
Joined
Oct 7, 2000
Messages
25,270
I mostly opened this thread to see if others are having similar issues with this card here at [H].
yes and not just your model and most are having problems on corsair psus...
think what you want, but thats where i would look before rma.
 

dr.stevil

[H]F Junkie
Joined
Sep 26, 2008
Messages
9,143
I was experiencing some weird things with my 3090 FTW3 from time to time (occasional flickering image and/or black screen crash). I did a bunch of reading on these issues and, while there are a lot of theories floating around as to what’s causing it, it “seems” like it may be related to the way the GPU is regulating voltage. Either drawing too much from the pci-e slot and/or causing voltage spikes through the 8pins. I can’t say for certain either way, but I did undervolt the card for a variety of reasons and knock on wood, no more flickering or black screens.

if you have issues with your RMA card (it sounds like you probably will based on feedback @ the EVGA forums), give undervolting a shot until EVGA is able to properly address the problem
 

Canon

2[H]4U
Joined
Aug 12, 2004
Messages
2,561
EVGA has issued an RMA and I will be shipping the card back to them.

Not much I can do here since the card won't display anything with known good hardware at this point.

Guess I got burned on this purchase. Hopefully I will be lucky...
 

Falkentyne

[H]ard|Gawd
Joined
Jul 19, 2000
Messages
1,721
His card is dead.
His RAM can't be causing the issue if the card won't even POST.
His PSU is absolutely NOT causing the issue.
There's a difference between an unstable GPU due to RAM timings (yes this can happen) and a completely DEAD GPU. Dead GPU's don't happen because of RAM timings. And the "PSU OCP" issue with 3090's (e.g. with older Seasonic based platforms) doesn't kill video cards. It just makes the CPU hard shutdown as if the power were unplugged.

It's important to look at the symptoms.

A crash (driver crash or regular black screen), no fan speed change, can be various system instability issues--like daisy chained PSU cables, 12v dropping too low (bad/weak PSU or high resistance cables) or any number of things like unstable RAM, unstable CPU, cache, etc. or you overclocked your video card too far.

A 100% (often on some video cards this is actually PAST 100%) fan speed crash and black screen is a low level hardware failure on the card. This is the same thing that happens for example, if you forget to put thermal paste on your video card GPU or heatsink and the card thermal trips and shuts down, or you use too thick thermal pads and the GPU doesn't make contact with the cold plate. Power regulation failures can cause this exact same thing, out of spec voltages, failing VRM's etc.
 

pendragon1

Fully [H]
Joined
Oct 7, 2000
Messages
25,270
His card is dead.
His RAM can't be causing the issue if the card won't even POST.
His PSU is absolutely NOT causing the issue.
There's a difference between an unstable GPU due to RAM timings (yes this can happen) and a completely DEAD GPU. Dead GPU's don't happen because of RAM timings. And the "PSU OCP" issue with 3090's (e.g. with older Seasonic based platforms) doesn't kill video cards. It just makes the CPU hard shutdown as if the power were unplugged.

It's important to look at the symptoms.

A crash (driver crash or regular black screen), no fan speed change, can be various system instability issues--like daisy chained PSU cables, 12v dropping too low (bad/weak PSU or high resistance cables) or any number of things like unstable RAM, unstable CPU, cache, etc. or you overclocked your video card too far.

A 100% (often on some video cards this is actually PAST 100%) fan speed crash and black screen is a low level hardware failure on the card. This is the same thing that happens for example, if you forget to put thermal paste on your video card GPU or heatsink and the card thermal trips and shuts down, or you use too thick thermal pads and the GPU doesn't make contact with the cold plate. Power regulation failures can cause this exact same thing, out of spec voltages, failing VRM's etc.
card works and goes black during gaming, system is still running and can shutdown morally via power button. so it does post.
others have argued about the psu...
yeah maybe to everything else.
 

Canon

2[H]4U
Joined
Aug 12, 2004
Messages
2,561
There seems to be some confusion in this thread already, allow me to clarify...

The issue started with the system still functioning with 100% fan speeds and a black screen which was recoverable by a system restart. Also sometimes instead of a shutdown it would just ramp the fans and settle down after a few minutes. That is no longer the case, the GPU will not display video at all, not even the BIOS post screens, no output whatsoever from any of the DP outputs on the card. My RTX 2080 Ti has been working perfectly since swapping it back into my system.

For those asking my card is made in China and has the "red lips".

As far as PSU... I have a new HX850 that I am using. It is set to single +12v rail mode and I have run an individual PCI-E cable for each of the 3x 8 pin PCI-E connectors. I also used this card with a Corsair RM 750 PSU with 2x PCI-E cables, one daisy chained because that is all this PSU has. The issue was the same on both PSUs. To echo what Falkentyne was saying, this isn't an OCP PSU issue, the system is not doing a hard shutdown or restart when these issues occur.

Anyhow, EVGA support was fairly responsive and generated an RMA for me and my card is shipping out today. I'm pretty disappointed about the RMA destination being in California USA... with me being in Canada. I am praying that this is not affecting all or most of these cards, this is pretty much why I posted this thread. I was curious if anyone here has had the same thing happen to them. So far no one in this thread seems to have experienced the issue.
 
Last edited:

Canon

2[H]4U
Joined
Aug 12, 2004
Messages
2,561
No bios flashing, no overclockig, no modifications, I was literally just using the card as it came out of the box. No point in doing any of that on an i7 6700k system, LOL. I've been waiting a couple months for a processor to finish the build that I literally planned around this card...
 

pendragon1

Fully [H]
Joined
Oct 7, 2000
Messages
25,270
There seems to be some confusion in this thread already, allow me to clarify...

The issue started with the system still functioning with 100% fan speeds and a black screen which was recoverable by a system restart. Also sometimes instead of a shutdown it would just ramp the fans and settle down after a few minutes. That is no longer the case, the GPU will not display video at all, not even the BIOS post screens, no output whatsoever from any of the DP outputs on the card. My RTX 2080 Ti has been working perfectly since swapping it back into my system.

For those asking my card is made in China and has the "red lips".

As far as PSU... I have a new HX850 that I am using. It is set to single +12v rail mode and I have run an individual PCI-E cable for each of the 3x 8 pin PCI-E connectors. I also used this card with a Corsair RM 750 PSU with 2x PCI-E cables, one daisy chained because that is all this PSU has. The issue was the same on both PSUs. To echo what Falkentyne was saying, this isn't an OCP PSU issue, the system is not doing a hard shutdown or restart when these issues occur.

Anyhow, EVGA support was fairly responsive and generated an RMA for me and my card is shipping out today. I'm pretty disappointed about the RMA destination being in California USA... with me being in Canada. I am praying that this is not affecting all or most of these cards, this is pretty much why I posted this thread. I was curious if anyone here has had the same thing happen to them. So far no one in this thread seems to have experienced the issue.
no shit there is confusion, you left out the part where its not working at all anymore. hope the rma works out.
 

Crotan

Gawd
Joined
Jan 2, 2007
Messages
935
For those asking my card is made in China and has the "red lips".
lol thanks for this. I've been gaming on a 3090 FTW3 Ultra since November that has "red lips". Never knew such a distinction existed. And when looking this up I saw the evga forums posts where someone put googly eyes over those lips.

I flashed the evga beta bios for a little while that allowed you to take the power target to 119% in X1 but eventually returned to the regular OC FTW3 Ultra bios, because I saw no discernable difference.

I realize you've already RMAed it, but I'm curious if you tried flipping the switch on the card to try and boot from the other BIOS? Since they have the OC and normal bios on at least the FTW3's
 
Top