The system experienced a crash, monitors shut down, the PC became unresponsive, and I needed to perform a power cycle.
The system experienced a crash, monitors shut down, the PC became unresponsive, and I needed to perform a power cycle.
Hi All
I'm getting desperate, as I'm not able to narrow down what is the actual issue.
My PC randomly turns of my monitors (dual monitor, second monitor is used for browsing/filming and is connected to a docking station for my work laptop
The crash itself is not a crash, just the monitors turn off instantly, due to no signal, but music/film/netflix/even the game is activelly running still.
When this happens, the PC ventillators go to 100% (likely the housing/CPU vents by the sounds of it) and stays like that until i manually power cycle (left it there one time for 10+ minutes, no difference)
Specs:
GeForce RTX 4070 Ti Trinity
i7-13700K
Kingston Fury Beast Memory, DDR5, 32GB, 5600MHz, CL36
Prime Z790-P
DeepCool AK620 CPU Cooler
Gigabyte P850GM 850W Power Supply
Versa J24 TG RGB housing
2 WD SSD and 1 Lexar NM620 1TB M.2 2280 PCI-E x4 Gen3 NVMe SSD
OS: Win 10 v 10.0.19045 running on of the SSD (have 11 installed on the NVME, but i dispise 11 so much im still using 10)
No overclocks, turned off in BIOS as well, BIOS was update less then a year ago (not latest, but have no access to pendrive for BIOS update)
Issue started right after Christmas, while playing PoE2, had 2 crashes in one night (i was like okay, sure, i was messing with graphics, namely downgrading as much as i could to get more stable FPS)
Then started playing ASKA, 3 days later, bumm, another monitor shutdown.
Then was just filming, browsing, playing some Captain of Industry, no probs, 5 days later, another crash.
Then 5 days later another crash.
Meanwhile i did some stress testing (3Dmark, OCCT for CPU, GPU, combined, memory and power supply) never had problem.
Did windows memory test, no fault found.
Updated graphics driver, several times, did a clean install with latest 591.74, now a 2nd time, in hope it is really a clean install.
Ran SFC /scannow and Dism /Online /Cleanup-Image /RestoreHealth, it did found some errors and corrected it, but 5 days later got my latest crash.
I pulled the PC apart, removed the RTX, de-dusted the fans, unplugged the power cord on the RTX then reassembled, 2 weeks ago, no change.
Housing is under my desk, but it is basically totally open, as my desk is a flat desk with metal legs, so airflow is even better than before when the housing was inside my old desk in a semi enclosed shelf (backend was close to a corner wall + bed, so 10 times worse airflow)
I've went through countless topics and forums but found no solution, as it was stated it could be CPU, GPU, memory, power supply issues.
Went through event logs and reliability logs, the only thing i found is this:
Description
A problem with your hardware caused Windows to stop working correctly.
Problem signature
Problem Event Name: LiveKernelEvent
Code: 141
Parameter 1: ffffdf87d14bf460
Parameter 2: fffff8046c17cd90
Parameter 3: 0
Parameter 4: 0
OS version: 10_0_19045
Service Pack: 0_0
Product: 256_1
OS Version: 10.0.19045.2.0.0.256.48
Locale ID: 1033
Had the 141 error every time when crashed happened, followed by the windows did not shutdown properly message (due to power cycle, at least if i can trust the timings)
I bought this setup 3 years ago in feb, so most of my parts are still on warranty, but some will expire this feb, so would be nice to know what is the culprit here.
If anyone had similar issues or any idea what can i do, please enlighten me, cause this is getting on my nerves.
The only thing i can do is to do a clean install of 11, but i would like to avoid that if possible at least for now, due to how much i hate that bloatware.
And due to reaching end of warranty, dont want to mess with the hardware too much not to make the warranty void if possible.
My guess would go for GPU as the most likely candidate, but due to how rarely the issue happens i cant really reproduce it. As a 2nd guess maybe the thermal pastes gave away (either CPU or GPU) due to 3 years of using, but then again, i never overclocked and i prefer not to stress my system to the max (usually turning graphics lower for more stable FPS and my system runs on 1080p all the time 144 Hz max (games included)
Thanks
p.s.
Even under heavy loads in games the temp stays mid range (CPU momentary peaks are 90 C (avg is ~70) for a second or so, GPU is also below 80
While i did stress tests, CPU peaks reached 100C (kinda expected) and GPU hit 85C as highest during 100% stress test.
But even when the system crashed, i checked the air it pushed out and it was lukewarm at best, so dont think it is overheating.
UPDATE
I did another stress test, all passed.
10 mins later, another crash while browsing, absolutely no load whatsover on the system. (first time ever, all previous was while game running in 1 screen, movie on the 2nd)
For relevance, please note the BIOS version on your motherboard.
Gigabyte P850GM 850W Power Supply
This PSU seems outdated; reviewing the details might help.
The device you own is described as problematic, I wouldn't stick with it.
I purchased this system three years ago in February, so many components are still under warranty, though some will end this February—knowing the cause would be useful.
Consider obtaining a more reliable PSU capable of delivering 650W to see if the problem remains. If not, you've identified the issue.
BIOS Version 1805
PSU is also three years old, all the PC components were purchased at once, except for the two SSDs. I plan to verify the S/N number of the PSU; I hope it’s an updated model since I bought it about 1.5 years after the announcement, but I can’t be sure how long it was sitting unused (everything came from the same webshop).
This is the first real clue that might help, so thank you very much.
My brother’s PC has a perfectly normal PSU, but because it’s known to be a problematic model, I’ll take it to the service near my block. They should test the PSU properly or replace it temporarily.
Do you have any suggestions for a replacement unit? I’m surprised you recommended 650W, as when I set up the system I was told to choose at least 800W just to be safe.
Also, since I’m not very technical, if the PSU is faulty, could it mean the system needs a short power burst and the PSU can’t handle it? That would explain why the fans spin at full speed?
Do you have any suggestions for a replacement unit?
Where are you situated? What is your budget for purchasing a PSU? Which online store do you prefer for buying?
A system that restarts unexpectedly may occur due to one of two causes:
1) an overheating problem
or
2) the PSU cannot supply power to the whole system.
Yet why would the ventillators run at maximum speed?
A reboot occurs but there is no POST.
GeForce RTX 4070 Ti Trinity
https://www.techpowerup.com/gpu-specs/zo...ity.b10293
I was a bit taken aback when you recommended 650W, especially since after setting up I was advised to choose at least 800W for safety.
I mentioned 650W with a little extra margin... I would have opted for an 850W unit if I had built the system myself.
Poland, I've purchased everything from the site you mentioned. I bought the current PSU for around 100 dollars, so I'm keeping it in the same range. If more is needed, I'm okay with paying a bit extra for a better PSU to safeguard my hardware rather than risking damage over time.
I don't think overheating is a big concern. Whenever the system crashed, I checked right away and saw the fans blowing air out, which was almost cool. Even the side panel felt warm but not hot when I touched it. The fans started from low speed (20-30%) and went straight to full power (100%), so I don't believe it's overheating.
The system is functioning normally enough to handle the occasional random crash after several days of regular use. During those crashes, it becomes unresponsive with no visual output, but the audio still works perfectly—streaming or local videos play without issues, and background sounds are unaffected. The only solution I've tried is a manual power cycle. After that, the system works fine for a few days until another crash occurs.
Unusual malfunctions frequently stem from a faulty power supply unit. Unless your nearby shop offers a detailed dynamic psu tester priced around fifty thousand dollars, basic checks won't suffice. The standard approach involves swapping the psu with a reliable, high-quality replacement that meets adequate power requirements.
Yes, that won't happen, the shop is mainly a garage store so it's unlikely they can handle that kind of inspection. Still, I'll check if they can offer a PSU replacement, preferably one that's proven to work. For replacements, I was considering the Thermaltake Toughpower GF 850W.
UPDATE
Was sick as a dog for the week, couldnt really do any testing. Finally had the power to do some testing (sadly corner PC store is closed so could not test it there)
Installed a brand new win 11 (to the M2 drive, in case my 2.5'' SSD being faulty due to age after 6 years), finished updates and drivers and then started a game
Thanks to Nvidia app, it scaled the settings to the max at 4k, system crashed after finished loading a saved game.
An extra bit of info
With the fresh install Win 11 (24h2) when i checked reliability history, i got not only livekernerevent 141 errors but also LiveKernelEvent 1b8
Got that also when i tried a 3d rendering test, when GPU load was not even reaching 90% same crash.
From what i gathered, these 2 codes together heavily points to PSU issues.
I will be able to grab my brothers 4070 Ti super on Tuesday, will swap around the cards for testing
If i get a crash during any GPU load, then as you guys said, most likely faulty PSU.
If no error while heavy load, it is faulty card.
Sadly he has the very same PSU so not really sure if that is worth swapping around as well...
My only concern is if it is really PSU issue, will it damage my brothers GPU?
Or more likely this is an issue with PSU not able to provide ample peak power ?
Also if I would test the CPU to the max, wouldnt it be able to compare level of spike demands?
In which case i can just test max CPU load and if crash, a clear indication on PSU being the culprit here?
I have my hand on a brand new PSU.
After installation, the problem persists and is worsening rapidly.
Even basic stress tests are causing issues—sometimes just having the browser in the background or playing a video file triggers it.
It seems like the device might be overheating; during maximum load, the CPU reaches 90-100°C and cools down almost immediately to below 40°C when the load stops.
OR
There could be a problem with the PCI-e interface itself.
I believe I've identified the problem if anyone is interested.
I switched the BIOS PCIe configurations from auto to Gen3.
Testing showed no issues.
Then I changed it to Gen4 and ran a 45-minute stress test (full power, dynamic 3D at 70-100%).
Still no errors detected.
Now my 4070 Ti is PCIe 4.0, which should be stable, especially since I've had this system for three years and the latest BIOS update was applied properly after the problem began.
It might be a physical issue with the slot—specifically the first PCIe slot (PCIe 5.0 x16 SafeSlot Core+)—or possibly the drivers have been interfering with PCIe control after three years.
Using the same BIOS version for a year before the issue and having multiple graphics drivers that are mostly up to date seems to be the cause.
Changing from auto to Gen3/4 appears to resolve the matter.