Hello EVGA forum members. Long time reader, first time poster. I come seeking assistance/guidance with my EVGA 1080ti FTW3 Hybrid.

I had thought my system stable, and last week I would have told you that. But I have been restricting my GPU to 1080p gaming, due to my monitors.

I recently added an ultrawide display to my desk, and with the new resolution, 3440x1440, I am finding that my 1080ti is not at all stable. Games are consistently crashing to desktop.

specs:
EVGA x99 Micro2 mobo (current bios)
i7 6800k (stock)
EVGA 1080ti FTW3 Hybrid (stock)
4x Gskill 16GB DDR4-3200 (default @2400)
Samsung 970 Pro NVME 512 GB SSD (C drive)
Corsair HX750w PSU (older, purchased 2012)
2x Asus 1080p @ 60hz (only using 1 currently)
1x LG 3440x1440 @ 144hz (34GN850-B)


The card is connected with independent PCI-E power cables, in the 16x slot. The GPU rad is mounted vertically in the back, as the CPU cooler takes up the case upper position, but I leave the side panel off most of the time. Windows 10 was installed fresh not long ago.

CoD Warzone is what my group currently plays, and I have been fortunate in that I've not seen my system crash at 1080p settings. However, at current factory clock settings and 3440x1440 resolution, it will consistently crash to Windows with a 'dev error 6068'. I tried all the recommended fixes, but only restricting the card allows stable play.

The Nvidia 'debug mode' allows the game to launch, but it will still crash after some time. It's usually when the GPU and ram temps reach 60c. I've found that currently I must instead lower my target power to 80% AND turn on the shroud fan manually, in order to reach stable play. If I keep the temps in the 50c range then the card is usable. Im using Precision X1 on the 2nd display to monitor and make adjustments.

I've been seeing similar instability in Civ6, Mirrors Edge, and Conarium.

I've got some Noctua thermal compound on order, and I am thinking about removing the cooler and reinstalling with a new coating. Beyond that, I might try moving the card into the next PCI-E slot, an 8x.

Thoughts? Advice? I had first wondered if the PSU might be at fault due to age, but after searching this forum for similar reports, I am starting to think it may be a bad card.

This card was purchased direct from EVGA and the warranty is valid for a few more months.

Thanks for the help!

Precision X1 or XOC (older version)  ?
 
either way slide Temp & Power Targets full Max
 
Slide Voltage to Max
 
Set GPU & RAM at "+0"
 
Case fans Max RPM
 
 
Temps ? 
 
Room
GPU
RAM
 
Try running Time Spy Benchmark

Any Free Benchmark .. is better than a Game for Testing
 
3DMark usually has a Sale from time to time - in the past < $5 got the basic non-profession benchmarks
 
 
- Benchmark is consistent load
 
- Games can have bugs & driver issues
 
- online games can have may different internet issues Or the service like Steam can has issues
 
--> only downside to benchmark software --> it is a synthetic load
 
Removing any Overclock on a system is always a good place to start troubleshooting

I'm getting a very similar issue (1080ti SC BE) - 3440x1440 monitor. 
This was a card I recently received from RMA, after my other 1080ti was crashing similarly, was replaced with this one.  
Everything worked well for about 1 month until the crashes started happening. First in Destiny 2 (broccoli error), then CD:MW 6068, and FS2020 Alpha (CTD)... 
 
If I run at lower resolutions, I can play for longer but still received crashes shortly after starting a game (sometimes even in menus)
Strangely, If I jack the fan to 100% temps stay below 65C, but still the crashes come, I guess not surprisingly since I get crashes at idle temperatures. 
 
PSU: Seasonic GX-850
CPU: 3700x
MOBO: ROG Crosshair VIII Hero wifi
RAM: G.Skill Trident 3600 32GB
Noctua coolers all round NHD15 on the cpu. 
 
Hopefully I can help troubleshoot along with RedCuba
 

RedCuba
A99Charlie: I'm sorry to hear that you are having similar trouble. I am able to game if I throttle to 80% power target. Maybe give that a try.
 
As for my own situation, I just finished transplanting the card into my friends computer.  Set max for power target, voltage and fan speed. Time Spy continues to crash within the first 20 seconds. Only after I set power target of 80% did 3dMark complete with a score in the 8000 range.
 
I will continue to tinker with this card, but it seems it's time to consider an RMA.
 
Thanks again everyone for their input.

RedCuba, Thank you very much for this suggestion... I was able to play MW and Destiny 2 for about 2 hours without a crash. Though at 1440p there is noticable drop in frames at higher settings. So I'm guessing the next step is to tweak the power settings to gain a little more performance before RMA? or just RMA?

Cool GTX
looks like RMA Time
 
Could be a part on the PCB .. but
 
If you like to work on your cards ... removing cooler ... You are allowed to replace the TIM & thermal pads on a EVGA GPU and Not Void your warranty --> unless you cause physical damage tot he card
 
 
I'm guessing your Temps go much higher when you run over 80% power

Thank you for confirming the info about not voiding the warranty. I had originally considered doing as you suggested with the cooler. I purchased a large tube of Noctua NT-H2 for the purpose of cleaning up and replacing the TIM.

I also found a tear-down video: search 'EVGA 1080 Ti FTW3 Hybrid Tear-Down' on youtube
Dissassembly looks straight forward. @9:10 it shows the TIM layer between the VRM plate and GPU cooler.

I know that thermal compound can dry up and need replacement, but do thermal pads wear out too? What would that look like? Dry and brittle? Discoloration?

But if this was truly an issue with overheating, wouldn't at least one of the ICX sensors indicate dangerous values? Even when the card is running at factory clock speeds, I don't think I have seen any of the sensors hit more than 60c. Also, the Time Spy crashing @3440 happens very rapidly, and from a cold GPU state. There isn't much time for the components to heat up significantly.

I have contacted EVGA and begun the RMA process.
 
 

Thanks for the input, arestavo.
 
Welp, I couldn't let it go. I just couldn't return the card without taking a look at the thermal paste. Pictures are attached.
 
The TIM seems to be making reduced contact; maybe a bit dry as well. I will clean and reinstall tomorrow.

Nice. 👍