cancel
Showing results for 
Search instead for 
Did you mean: 

TUF 3090 Ti O24G - Crashing to Black Screens

iamrizzo
Level 7
Motherboard model: ROG STRIX Z690-E GAMING WIFI

UEFI Version: 1505

CPU: 129000KS

Memory kit part number and amount in GB: KF552C40BBAK2 64GB - Kingston Fury Beast 5200MT/s

GPU: ASUS TUF 3090 Ti (O24G) Gaming

SSD/HDDs/Optical drives: 2x Samsung 980 Pro M.2 2TB

PSU: Corsair HX 1200

USB Devices (model/version number):

Ducky Shine 7 Mechanical Keyboard
Logitech G9x Mouse
Shure PG27USB Microphone


CPU Cooler: AIO Arctic Liquid Freezer II 420mm

PC CASE: Fractal Design Torrent

Operating system: Windows 11 64 bit
Microsoft Activated yes/no? YES


Drivers Installed (include version info):

Intel Net: 1.1.3.28
Intel net: 22.170.0.3
Intel System: 10.1.45.4
Intel Management Engine WMI Provider: 2130.1.15.0
Intel Management Interface #1: 2131.1.4.0
Intel Dynamic Application Loader Host Interface: 1.41.2021.121
Intel Wireless Bluetooth(R): 22.150.0.6
Intel Ethernet Controller (3) I225-V: 1.1.3.28
Intel Wi-Fi 6E AX211 160MHz: 22.170.0.3
Intel USB 3.20 eXtensible Host Controller - 1.20 (Microsoft): 10.0.22000.778

Nvidia Display Driver: 31.0.15.1659
Nvidia Game Ready Driver: 516.59
Nvidia High Def. Audio: 1.3.39.3
Nvidia Virtual Audio Device (Wave Extensible) (WDM): 4.45.0.0
Windows Audio Driver: 10.0.22000.1

AOC 2757: 1.0.0.0
Generic PnP Monitor: 10.0.22000.1042


Any third Party temp/voltage software installed: No

System Overclocked (provide details)?: No

CPU Set to default settings in Bios
XMP disabled currently in Bios, however having XMP enabled or disabled doesn't change the result
Card is at default clock from the factory



Issue:

Experiencing an interesting type of crash I believe is coming from the GPU. When running video games or other graphically intensive applications like Heaven Benchmark 4.0, Red Dead Redemption 2, Blender Renders, etc. This crash causes both of my monitors to go black, and into 'inactive' mode. Audio will continue playing through my headphones, but I always have to press the power button to let windows shut down and reset the driver before I can get video back. Also on a rare occasion, when booting the PC after a crash, the windows login screen will display very odd artifacting, which only a quick glance at the bios (don't need to change anything) will resolve.

When the driver crashes, the GPU will display a red LED light indication next to the 12pin Nvidia PCIE power cable adapter, which I believe indicates that no power is coming into the unit? The same Red LED turns on when the system is off, and stays on until the system is booted. When the GPU crashes and the LED turns on, about 30-60 seconds later the LED will turn back off, but the display drivers do not recover (this is before I press the power button to reboot the system).

I have taken a video of the crashing process which I can upload, running Heaven Benchmark 4.0 and monitoring HWinfo. To my knowledge, there is not any temperatures out of line, nor is there appear to be any abnormal voltages.

I have taken the following steps in my own troubleshooting process:
- Fresh install Windows 11 (Excluding Armoury Crate and other similar softwares)
- Set everything to default values in BIOS
- Reseat GPU into PCIe Lane
- Reseat all power cables between GPU and PSU
- Roll back Nvidia Driver from 522.25 to 516.59
- Roll back BIOS from 1720 to 1505
- Turn on/off XMP Profiles
- Reseat RAM
- Move M.2 drive off of the shared Gen5 lane with GPU
- Swapped GPU out for an old EVGA 1050ti and ran HeavenBenchmark4 with no crash

Thanks for coming to my ted talk, am happy to discuss and troubleshoot further, and provide more screenshots. I will be available to do so. Not sure if I have a Lemon GPU on my hands and need to RMA or if there is some deeper compatibility issues with the rig I have built.
Someday, when I have a lot more time and money, I will become an overclocking God.
2,548 Views
9 REPLIES 9

iamrizzo
Level 7
Went ahead and ran a Cinebench test, and a Blender benchmark; cinebench completed and scored higher than the first time I ran it.

During blender I got black screen failure once it got to the classroom test both times I ran it. The first time, my graphics card recovered the driver and the desktop came back after a few seconds of black screen, it received a score of around 1500. Second time the entire PC rebooted and brought me back to login screen.

In event viewer, I see that there is a critical error: Kernel-Power, Event ID 41, Task Category 63.

A little bit of research leads me to believe that the power supply is the culprit? A question I have regarding this would be why does the PC not reboot after every time this happens? I noticed that the red LED next to the 12pin power adapter didn't light up during this crash, or it happened so fast I didn't see it. I would imagine if overcurrent or something was being triggered it would always shut down the whole system? And the event viewer error could point to a driver issue or a bios power setting as well, correct? Keeping in mind that the only reason I am even able to see this event happen is because the system reset, rather than the regular driver crash which gives results of no event being recorded.


Here is a list of the applications that have been related to a crash:
- Red Dead Redemption 2
- Max Payne 3
- Apex Legends
- No Mans Sky
- Overwatch 2
- Battlefield V
- DIRT Rally 2.0
- Heaven Benchmark 4.0
- Blender Classroom
- GPU Tweak 3 - OC Scanner
Someday, when I have a lot more time and money, I will become an overclocking God.

Attached a screenshot from a video I took running the Heaven Benchmark 4.0 showing data from HWinfo. This is the last frame before the monitors go black.

Anyone spot anything odd?
Someday, when I have a lot more time and money, I will become an overclocking God.

Was poking around after downloading GPU Tweak III to see if I could spot any potential issues, found that when I attempted to run an OC Scan, it would repeatedly fail the process at around 21%. Not sure if it's something I am doing wrong, or if there is an issue with voltages coming through the card. Tried to set my card into 'Q' mode compared to 'P' mode but no differences were made.

Attached a log file below..
Someday, when I have a lot more time and money, I will become an overclocking God.

Went ahead and frustratingly purchased a different power supply I found at a local computer store, as Asus Tech Support advised me that they couldn't help me unless I did so. It was also an upgrade in terms of total possible wattage being a Corsair HX1500i compared to the HX1200 that is currently installed in the system.

The 3090 ended up crashing out on me in the same manner, at random intervals of time after several benchmark tests at different GPU power demands.

At this point it seems pretty clear to me that the GPU is the issue, and I am awaiting Asus Customer Support to return my email to being an RMA process.

I also attached the most recent log from the Corsair iCUE logging to try and catch anything suspicious, if there are any professional log chart viewers here, would definitely love the help.

Lastly, attached a screenshot of the only thing I could find that seemed odd. At the moment of the crash, it looks like MB Temperature Sensor #17 is reporting a temp of 109C? Not sure what to make of that.
Someday, when I have a lot more time and money, I will become an overclocking God.

iamrizzo wrote:
Was poking around after downloading GPU Tweak III to see if I could spot any potential issues, found that when I attempted to run an OC Scan, it would repeatedly fail the process at around 21%. Not sure if it's something I am doing wrong, or if there is an issue with voltages coming through the card. Tried to set my card into 'Q' mode compared to 'P' mode but no differences were made.

Attached a log file below..


have you checked what the event logs are saying is happening around the time of the crash?

Soupladel wrote:
have you checked what the event logs are saying is happening around the time of the crash?


Hey, yeah I did check the event logs in windows, as well as did my best to track down any possible outliers in the charts of the logs from HWiNFO, iCUE, GPU-Z, with no luck. Unfortunately the card doesn't register an event happening when it crashes...
Someday, when I have a lot more time and money, I will become an overclocking God.

Alright, so after much back and forth with Asus tech support. I decided to just refund the card and get something a bit more stable.

After doing another round of extensive research, and without opening the card up and voiding the warranty, I have landed on very poor thermals happening within the heatsink itself. Whether it is because of a lack of proper thermal pad application, or because the TUF 3090 Ti heatsink has an insufficient design, the card would run up to the high 70's and low 80's pretty quickly after starting most graphics heavy applications. And to my knowledge, running any higher than 83C on these things will cause all sorts of issues to become present. (correct me if I'm wrong)


I did find an interesting youtube video that discussed the thermals of the TUF 3090ti, and it suggests that the issue at hand is that the card has been given a potentially insufficient heatsink...? Like the heatsink from the TUF 3070ti and 3080 were just basically slapped onto the 3090ti OC editions? Seems weird.

Link here: https://www.youtube.com/watch?v=2peZ_BN6q6A


I am honestly quite disappointed with the TUF graphics card itself, and with Asus tech support. On the other hand I am quite happy with the performance of my ROG motherboard, and look forward to using it more. So I am going to chalk this up to being the card was built on a Monday, and while I have read reviews about people having okay experiences with the TUF 3090ti O24G, no review I have found was speaking on exceptional terms about the card, all mentioned loud fan noise, warm thermals, throttling, etc.

I know there isn't much response to be garnered here, but hopefully this thread can serve some purpose to someone in the future who may be having thermal/black screen crash issues with their Asus TUF card.


If that's you, and you can't claim a warranty on it, time to get a water block.



Signing off,
Someday, when I have a lot more time and money, I will become an overclocking God.

This is probably unrelated, but on a Z390 system last year I was having very similar black screen issues. I tried just about everything - BIOS, motherboard and memory settings, several reinstalls of Windows, replacement PSU, different mouse, keyboard. I could go on. I spent weeks going through every permutation of settings.

In the end, my black screens were being caused by a defective USB cable. Amazingly, once I replaced that, all was good.

Like I said, probably unrelated. I have the same GPU and personally I've found it very good. It always runs cool, makes no noise and has been solid. I use MSI Afterburner to limit temp and use a custom fan curve. I can limit it to 65C and it runs every game I have without problems. Maybe I was lucky, but the TUF 3090TI has been good for me (so far, lol). I hope you are happy with your new card. What did you get?
Z690 Hero, BIOS 3401, MEI 2345.5.3.0, ME Firmware 16.1.30.2361, 7000X Case, RM1000x PSU, i9 12900K, ASUS TUF OC 3090TI, 2 x 16GB Corsair RAM @ 5200MHz, Windows 11 Pro 23H2, Corsair H150i Elite AIO, 4x Corsair RGB fans, 3x M.2 NVME drives, 2x SATA SSDs, 2x SATA HDs.

Hey John,

Interesting, I hadn't swapped out any cables besides the display cables that I have from my GPU to my monitors.

What really showed me that the card was having thermal issues, is right next to where my computer lives I have a window pretty much right above it; and one night it got a bit colder than usual, so I decided to open up the window and let my PC intake ~50 degree air right into the chassis, and the problem went dissappeared. Given that I currently live in an area that tends to get a bit hotter than average during the summer months, constantly providing 50 degree air into my chassis was going to become an impossible task lol.

I ended up going with an EVGA 3090ti FTW3 Ultra, which spec wise is almost identical to the TUF 3090ti with a bit larger of a heatsink on the card, and it worked first time and hasn't given me any crashes or issues since.
Someday, when I have a lot more time and money, I will become an overclocking God.