New PSU and GPU - computer is unstable with random freezes, crashes, and loss of display?
New PSU and GPU - computer is unstable with random freezes, crashes, and loss of display?
Problem is as follows – I’ll try to keep it as short as possible.
First – my build:
Original:
Gigabyte Z490 Gaming X AX v. 1.00 BIOS v. F3
32GB Ram dual-channel (2x16 GB) DDR4 Crucial BL16G36C16U XMP profile 1
Intel Core i9-10850K with custom cooling tower (Noctua NH something, I don’t remember exactly – dual fan)
Gigabyte Nvidia RTX 3060Ti Windforce OC
Modecom Volcano Gold 750W
3 drives: SSD Samsung EVO 1TB NVME (boot) and two, older HDDs
2 screens (main DP 1440x2560 and additional HDMI 1080)
Windows 10 64 bit
no user OC, everything stock, as it were in the box. only XMP profile active
Lian Li Lancool II case with two fans at the front and one at the back
That build had ~4 years. General de-dusting was done around may 2025, but thermal paste on both CPU and GPU were not changed then (on CPU it was replaced like a year, year and a half ago; on the GPU this was never done).
New:
Gigabyte Z490 Gaming X AX v. 1.00
BIOS v. F22
32GB Ram dual-channel (2x16 GB) DDR4 Crucial BL16G36C16U XMP profile 1
Intel Core i9-10850K with custom cooling tower (Noctua NH something, I don’t remember exactly – dual fan)
Gigabyte Nvidia RTX 5070Ti Windforce OC
BeQuiet Power Zone 2 1000W 80 plus platinum
3 drives: SSD Samsung EVO 1TB NVME (boot) and two, older HDDs
2 screens (main DP 1440x2560 and additional HDMI 1080)
Windows 10 64 bit
no user OC, everything stock, as it were in the box. only XMP profile active
Lian Li Lancool II case with two fans at the front and one at the back
Prelude:
Last week during YT watching – suddenly, with no prior warning the displays blanked and stopped getting signal from the GPU and her fans got up to 100% speed. I still heard video playback perfectly fine in the background, so the system itself was still running. I forced restart with a POWER button, but GPU didn’t produce any signal on any of the screens. MOBO was passing POST with VGA error (diode in the bottom corner) and Windows was booting as normal – only I couldn’t see that – only hear. Upon changing the HDMI to the integrated port in the MOBO – screen was up (only 1080p one).
I started to search for an answer – graphics card itself wasn’t smelling or look funny – but I didn’t dissasemble it, so I can’t tell for sure. The PCIe slot also looked fine. Card was firmly and snuggly mounted. Later I confirmed, by mounting it in another computer, that RTX 3060Ti does not display anything for 100% .
With that, as when called, a handful of repair shops told me, that they won’t even look at the card and fixing it is not wortwhile anyway, I had to buy something. It’s one of the worst times to buy GPU, but ultimately I’ve decided (maybe to my detriment) after a day of reading, watching and comparing, that I’ll at least get something out of the situation and get an upgrade – I’ll buy RTX 5070Ti. With that I decided to also replace the PSU, as I wasn’t sure, if it was 100% fine (maybe 3060 died because of it) + my 10+ year old Modecom had just the same 750W as the 5070 TDP listed, so I wanted a safe headroom. I bought 1000W BeQuiet, as in the new spec above.
I mounted them both, along with a stand for RTX (tightly, so the card couldn’t move or sag) and replacing all of the PSU cables. I connected the card with this fancy new 600W connector with additional pins (I used a cable supplied with the PSU, I didn’t use the adapter that came with the card).
Problem:
With the first power up, MOBO passed the POST with no errors and I breathed a big sigh of relief, as I saw main screen light up with a familliar logos. Windows booted with some defaut config with no display drivers and like 480x640 resolution. I clicked some and started to download Nvidia drivers... and during download everything died – display went black, fans were just slightly more aggrievated and the case POWER button was the only option. After restarting, I downloaded the drivers, installed them (the second screen got alive again too) and everything seemed happy. I tried Kingdom Come 2, to see, if it won’t collapse under load – so for around 5 minutes I ran through the forest and the nomads camp with no lag, stutter or any issue, so I thought it’s a success, right? Well, not quite, as after I closed Kingdom Come, and went to close the Steam – computer froze and hung. Just the button restart left at this point. So, since then (about 5 days) i diagnose, test and nothing is really working.
How things are now:
Randomly, after about 10 minutes to an hour with some change typically, everything hungs (display and sound freezes, nothing can be done, fans up in speed slightly). The rare occurence was, that the display went black (no signal) first, but the system itself was still working for several seconds, befor it also crashed. Once Windows managed to reboot. Doesn’t matter the load (YT video, game) or no load (just desktop). It hung when downloading MSI Afterburner, upon closing Steam, once or twice just like that on the desktop, when idle and it also hangs while gaming. Or when YT video plays. No rhyme or rhytm.
I used DDU in safe mode to wipe graphics drivers, and let Windows install them automatically. This seemed to help – for more, than two hours I played World Of Tanks without issues. Seeming to help stopped, when in one of the battles it started stutter and choke heavily, with FPS flying like crazy from 240 to 20 – 30 and back. It lasted for like a minute, before finally giving up and freezing. After that I downloaded HWiNFO and started logging sensors to a file. I played the next 40 minutes with no issue, until it froze again with no stutter or prior indication. I logged another 30 minutes of gameplay to the next freeze, though this time the display went black first, a couple of seconds before the system gave up. After that I thought, I’ll check it it is temperature and load related, so I played Worms Armageddon, which completely bore this PC build – for about 45 minutes and it froze again. I wanted to chceck the HWiNFO logs, but during browsing through one file, about 5 -10 minutes after last hang, (oops) it did it again. That was Saturday, night, so I leftit at that.
Sunday, first thing it the morning I played two battles in WOT, until the first freeze of the day. So from that point on, I started to try to solve this issue. Below I list everyhing I’ve done so far, every test, results and my observations:
- in Nvidia control panel I changed power management to [prefer maximum efficiency] – no dice;
- in BIOS I changed PCIe standard form [AUTO] to [GEN 3] – no luck still;
- I’ve combed throuh HWiNFO logs – to see, if there are any sudden or/and atypical voltages drops/risings, any thermal anomalies or other weirdness - nothing (or at least nothing I saw/ was logged/ is indicative, but I don’t know it’s important).
- HWiNfo shows things, that I can’t tell are normal or not (to me seem, like they’re not typical):
- in the GPU section, where some boolean indicators are shown, parameters [reached efficiency limit / power limit / voltage limit and reached reliability limit] there is a caleidoscope of [NO] and [YES] values changing near constanltly.
- also in the GPU section, under [PCIe Error Counters] I have loads of very fast accumulating errors. Those are [Replay Error count, Replay Rollover Error Count, Recovery Count, NAK's received and Recover count] – they go in hundreds, thousands and especially the last one – tens of thousands. With more load, like YT video or games they go up like crazy by tens a second, but on the desktop or even video playback from the hard drive they either slowly accumulate or stop at all;
- every MOLEX and connection on the MOBO was checked – new GPU is correctly mounted, sits in the PCIe slot correctly (as far, as I and my father can tell) and screwed solid, so I think, that everything around her will shatter, before she even flinches; all conections are tight and does not move, no cable is broken;
- in Windowses „event viewer” there is nothing, save for critical Kernel-Power "computer was not correctly shut down and restarted blah blah blah" – no driver failure, or other type of stuff;
- Windowses „reliability monitor” (who even knew this exists) there are some basic information about the crashes, and it gives „Hardware failure” with several error codes (that worries me some);
- little to no DUMP files are created, but yesterday I found, that windows managed to create one and upon checking – it lists display driver as the culprit, and as I read and watched JayzTwoCents and GamersNexus, the Nvidia drivers are junk in general – also Windows is kinda garbage – so maybe those are factors?
- We connected an additional power cable to the MOBO (only ATX 12V 2x4 was connected, we added also ATX 12V 2x2) – this changed the odds of display blanking first slightly and added more stutters just before the crash, but no significant change in the frequency of them;
- I updated BIOS to F22 (june 2023) [for some reason I can’t install F24a, as it says „incorrect file”] – this seemingly changed things – now there are random microfreezes / stutters with YT playback / in games – and they tent to lead to a crash. The crash itself is also changed – now almost always the screens loses signal first, and the system crashes some seconds later (but freezes with no warning and no load also occur, just significantly less often).
- I tried to use MSI Afterburner to limit the GPU power (one of my collegues had similliar problem and this was his fix, though his problem was random freezes under heavy load in „series” like – one frees once a year, then two or three a day for about a week, then several months freeze free) – I limited power to 90% and again – this resulted in a record (at the time) 2 hours and 38 minutes with no issues and then sudden freeze. I reverted back to 100%.
- for now, microfreezes and stutters just randomly happen – some time is ok, then it chokes for couple of seconds, if I stop the video / game then it goes away for some time, and sometimes it undergoes several of that cycles, sometimes it just sudden freezes or blanks (after a while). I also noticed audio clipping (characteristic hard "smack" sounds and robotic audio) and even USB peripheral problems – namely keyboard and mouse (mouse is a pass-throughed thorugh keyboard) litterally lags, hungs for a while or completely shuts down (I have to reconnect them to use them again) – this mostly happens during the „choking” but sometimes I can induse „the choking” by simply moving the cursor over the viedo playback! Despite this, I can play WOT or other game with minimal stutters once in the while (USB disabling still occurs). And then computer hung on the YT playback;
- I did update all of the MOBO drivers – chipset and all, despite the mess on gigabyte’s website – I did change nothing or very little;
- I enabled 4G decoding and resisable BAR in BIOS – nothing changed, only it lasted for about 4 hours, until it froze again (mixed use, some YT, some game, some internet). After that it froze again in just 10 minutes and again in another 5;
As I was sick and tired of it, I left the computer off (completely, unplugged from the wall) for about an hour, as I read some stuff on the phone. Then I wanted to test and see some more things I’ve read, so I turned computer back on – and this time it lasted thorough some tweaking, prodding and finally – YT for another like 1 and half to 2 hours – and it probably could last a little longer, but I turned it off to go to sleep. So letting all work capacities to decipate (by unplugging) seem to prolong the time to the next crash?
Questions
Does anyone have any idea what causes this (i know – hundreds of things, right?)? What (relatively) non-invasive tests can I do still, until unplugging, replacing and reinstalling all the things, including nuking windows installation and all the drivers?
For me it looks incresingly, like the motherboard. No driver error (or at least Windows does not report them) in event viewer (though in dump driver is listed), freezes happen with or without load at random, there is (seemingly) no reproducible conditions for hang to happen – just wait a bit or play video or a game and then wait a bit. Those PCIe errors are ramping up insanely fast. First to go is screen display (though it wasn’t the case earlier)... Is this reasoning in any way reasonable?
Question is also – what was first? Egg or a hen? Or a rooster?
Was it 3060, that damaged MOBO, or the MOBO, that killed 3060? Or it wasn’t related – 3060 just died and now I have other problem? I’ve read and watched some stuff, and generally a „mismatch” of components (like different PCIe generations) can bottlenech upper % of efficiency, but it still should work together stable.
Please, give me any idea or advice. At that point I consider buying new platform (MOBO, CPU) – thoug it’s easy to just buy stuff – my concern is, if I buy stuff and the problem persists still.
I have the HWiNFO logs and other things, if it helps. Take note, that I have them in my systems language (not english), but it should be easy to translate, if needed.
Your problems appear to stem from hardware rather than software, though more details will follow.
Most of these issues originate from the crucial part inside the PC. The one who supplies power, controls everything, and connects all devices is the PSU.
What? I’ve never heard of that brand before.
My understanding of PSUs is the most comprehensive of my knowledge about PC hardware.
I searched for the brand and found it to be a Polish in-house product, specializing in navigation tools (GPS), luggage, certain PC cases, and now a PSU?
Upon closer inspection of the PSU, the wattage on the rails seems accurate, but it lacks OVP and OTP safeguards. With these two missing, it raises doubts about the quality—possibly low or even subpar.
That clarifies things further. Notably, the PSU was included with a two-year warranty.
It seems you were testing your luck until the PSU failed completely, along with other components like your RTX 3060 Ti.
Additionally, there’s a high probability that the PSU damaged all other parts in your system—CPU, motherboard, RAM, SSD, etc.—by operating outside specifications, causing voltage issues, excessive ripple, or EMI. This gradual degradation can harm the whole setup over time rather than causing immediate failure.
Therefore, it’s reasonable to suspect that everything connected to that Modecom unit was faulty or degraded.
You’ve now swapped the PSU and GPU. What remains are the CPU, motherboard, and RAM. These three are essential for the PC to function correctly. An SSD or HDD is also necessary for the operating system to operate smoothly.
Even with a new PSU, it doesn’t automatically undo all previous damage.
By the way, your new Be Quiet! unit isn’t particularly impressive either. It’s of average quality, barely acceptable—around Tier B+.
For a gaming rig, a Tier A PSU is ideal. Examples include Seasonic Focus/Vertex/PRIME, Corsair RMx/RMi/HXi/AXi, Super Flower Leadex Gold/Platinum/Titanium.
PSU recommendations:
https://docs.google.com/spreadsheet...JW...1973454078
A Tier A unit is sufficient; Tier A+ is preferred.
(My three PCs are all powered by Tier A PSUs. Specifically, I have Seasonic PRIME 650 Titanium (Tier A+), Seasonic PRIME Ultra 650 Titanium (Tier A+), and Seasonic Focus PX-550 (Tier A). Detailed specs with photos are in my profile.)
Troubleshooting:
From a software standpoint:
Boot into Safe Mode (with networking) and attempt to freeze the PC.
In Safe Mode, gaming isn’t possible because only essential Windows drivers are loaded. However, since your PC also freezes when browsing the web (e.g., YouTube), try forcing a system freeze in Safe Mode.
If the system freezes in Safe Mode as well, the problem likely lies with the Windows kernel or hardware.
If it doesn’t freeze and everything works smoothly, the issue probably stems from driver or software conflicts.
To resolve these conflicts, there are two approaches:
* The lengthy and time-consuming method
* The quicker, simpler solution
But before I go into further details, please run a Safe Mode test. This will significantly narrow down the cause.
I value quick replies, thank you.
Yes, Modecom is a Polish firm and they manufacture PC hardware—actually, I thought they were a major tech brand like Asus or Dell, but I wasn’t even aware they originated from Poland until recently.
That’s not important now.
Following your advice, I ran the test in safe mode. The outcomes were:
Safe mode without network failed—Windows refused to boot completely, only the task bar appeared. Hovering over it showed a loading circle; clicking it displayed “windows don’t respond, wait or close.” When I closed it, the task bar reappeared and became functional again, though the desktop and all functions remained unresponsive.
However, the bare minimum safe mode without network worked. It felt quite basic, indicating no graphics adapter was detected, but it operated smoothly. After reviewing logs and settings, I recreated the situation and it froze only briefly—no freezes, crashes, or stutters for three hours now. The system appears stable and doesn’t show any abnormal fan behavior.
Good and unfavorable updates; this is positive because it shows the hardware and OS kernel at the lowest level are functioning properly. At least in the near future, as long as the obvious problem is resolved.
Unfavorable news. Yes, it’s not entirely negative since it has limited the possibilities significantly, suggesting a likely network-related problem.
However, it remains unfavorable because I lack experience with network problems and, based on my current searches, I can’t provide substantial assistance. (My expertise mainly covers hardware issues, though I have some understanding of software challenges as well.)
Here’s what I discovered:
If Safe Mode with Networking doesn’t activate, but regular Safe Mode does, the most probable reason is an issue with your network drivers or related services that aren’t loaded in standard Safe Mode.
To address this, you might try disabling the problematic network driver or service via Device Manager or using the System Configuration tool (msconfig) within Safe Mode.
If that doesn’t work, consider using System Restore, accessing recovery tools through Windows installation media, or performing deeper troubleshooting such as rebuilding your Boot Configuration Data (BCD).
If none of these steps succeed, you may need to reset everything by wiping the OS and installing a fresh, clean version of Windows. This will eliminate software-related problems, including bloatware and malware (except for firmware rootkits).
First, let’s try to get Safe Mode with Networking working again.
It begins simply:
1. Begin in standard Safe Mode:
- Press Shift + Restart and select 4) Enable Safe Mode.
2. Launch Device Manager:
- From within Safe Mode, right-click the Start button and choose Device Manager.
3. Inspect network drivers:
- Navigate to "Network adapters" and look for any devices marked with a yellow exclamation mark or a red cross, signaling trouble.
4. Turn off the problematic driver:
- Right-click the suspect network adapter and select Disable device. This should allow Safe Mode with Networking to launch next time. You can then try reinstalling or updating the driver once networking is restored.
5. Review system services:
- If the driver isn’t the culprit, run services.msc in a command prompt from Safe Mode and disable any unusual entries.
6. Use System Configuration (msconfig):
- In standard Safe Mode, type msconfig in the Start Menu search bar and open the System Configuration tool. On the Boot tab, ensure "Safe boot" is turned off, then restart to attempt a normal boot.
If these steps don’t resolve the issue, proceed with System Repair or System Restore.
If all else fails, you might need to format the OS drive and perform a fresh, clean Windows installation. This will remove all software-related complications, including bloatware and malware (except for firmware rootkits).
Initially, we should focus on getting Safe Mode with Networking back online.
It starts with ease:
I noticed the JayzTwoCents video from last week, where he found that wireless drivers (Bluetooth and wifi) led to numerous strange loading and unloading problems. This caused PCIe saturation, resulting in microstuttering and lagging games, plus random USB device disconnections and failures to reconnect. It was quite unusual and similar to other issues I’ve seen before. In the event viewer, I also observed some comparable errors. This likely explains a persistent bug I’ve had for years—random connection lag lasting up to a minute, losing signal, and struggling to establish a connection after 5 to 10 minutes, even with strong signal and range.
Currently, I’m in a stable "minimal" state (from msconfig diagnostic boot), and after about 30 minutes, things seem normal. No PCI errors are visible in HWiNFO (only minor Recovery Count).
I have no audio, no network, and just the basic graphics driver—no NVIDIA control panel either.
I was planning to take this PC to a repair shop, but I wanted to back up my data first so they could wipe and reinstall everything. Perhaps I’ll try reinstalling Windows cleanly now—it’s accumulated quite a bit over the years.
You can achieve this effortlessly with a second empty drive available.
The process involves:
* disconnecting all connected drives
* linking only the empty one
* installing the operating system onto that empty drive
* ensuring a successful boot to the OS
* reusing the previous OS drive as a backup, but avoid shutting it down
* after data recovery, format the old drive and clone the new OS
* optionally, retain the OS drive as a bootable backup
This method has proven useful in my data recovery work, especially for sharing with friends or family. It allows preserving information on the OS drive even when the operating system can no longer be accessed.
In my setup, I maintain one drive as a reliable backup. If the primary drive fails and the BIOS cannot start it, the system automatically switches to the second drive and loads the OS from there.
I don’t have an empty drive. It took me about an hour to go through the C:/ folder and collect what matters most—mainly my saved files, Firefox items, some clips, and screenshots. I moved them to the non-bootable HDDs, intending to just remove them before going to the shop (yes, I understand—they’re needed for configuration—to protect my family photos, important documents, and personal stuff—no one should access or delete it).
I ran a game in the diag boot that didn’t need a network and lasted nearly an hour without problems. Stable FPS, no stutters, no USB disconnections—this really suggests an exotic drivers issue. Tonight I’m finished. I plan to take early time off tomorrow and try reinstalling Windows to see if it helps. If not, I’ll go to the shop; they can test every part much faster and more thoroughly than I can. If something goes wrong, I should still have the option to return the card and PSU for a full refund (my 14-day return window ends next Wednesday).
Thanks a lot for your help so far. I’ll reach out tomorrow with any updates.
This was intended to be an upbeat update, but it’s only moderately positive at best.
The positive news is that turning off the on-board Wi-Fi adapter in Device Manager resolved the PCIe errors, random stutters, and freezes—lasting five hours without any issues. I played an offline game, watched two hours of YouTube, and explored the C:/ folder to decide what files to keep or delete.
However... when I enabled WOT, the system ran smoothly, audio stayed clear, and the situation improved significantly.
I experience some minor micro-stutters during battles, slight FPS reductions, and occasional audio pops (clipping) in certain moments. The main problem remains with input devices—both keyboard and mouse—becoming unresponsive or freezing unpredictably.
This might be connected to the World of Tanks network settings, possibly causing lag, which aligns with my past experience in that game. The fact that other tanks remain unaffected supports this theory.
Another possibility is interference from a WiFi extender, which has now become my Wi-Fi adapter (connected via Ethernet RJ). I can’t lay a full Ethernet cable, but if I could, I’d remember this situation from 15 years ago.
Performance-wise, the game loads slowly, textures take a long time to render, and it feels sluggish overall. It seems the game runs from one of the HDDs, which isn’t ideal for speed. Previously, this wasn’t the case.
Additionally, the temperatures reported in HWiNFO are concerning—my card heats up to around 55–65°C, which is acceptable, but the CPU reaches about 85°C. It’s becoming too hot for comfort.
I’m unsure what to do next. I’m torn between a full system reinstall and taking it to a repair shop.
Could this be because my GPU is overloaded? My system struggles because the GPU waits for the CPU to catch up, while the CPU barely functions. The PCIe bus becomes saturated, and the motherboard has trouble powering everything.
Also, disabling the on-board Wi-Fi adapter feels unsatisfactory. I rely on Wi-Fi for both regular connections and Oculus VR (for virtual desktops and certain games). Using an extender technically treats it as a cable connection, which isn’t ideal.
I’ll test Kingdom Come 2 to see if the input glitches stem from WOT connection issues or a heavy workload.
You now face two challenges. First, using games on HDDs is no longer advisable by 2025 due to their poor performance. Most new titles now list SSD storage as a necessary minimum. Consider reinstalling the game on an NVMe drive to check if this improves your experience. Second, playing online games over WiFi with an intermediary extender can be problematic. Although speed tests show acceptable results, interference from other electromagnetic waves often causes lag and data corruption. Even after the router requests resends of corrupted packets, this process adds delay. Using a low-quality extender may worsen the situation. You might try disabling the WiFi adapter entirely and uninstalling it via Device Manager, then restarting your system. This should prompt Windows to reinstall it, potentially resolving past issues. If this doesn’t help, switching to a better quality WiFi USB adapter could be beneficial.
By the way, be cautious when relying on Jaytwocents videos—his claims are often exaggerated and based on personal anecdotes rather than solid evidence.
I understand this topic well, especially regarding gaming over Wi-Fi. Unfortunately, I can't set up an Ethernet connection because of my living situation.
Yes, I plan to switch WOT to SSD eventually.
It's a bit strange, but it actually helped me.
There are other problems too, so after careful thought and some hesitation, I've chosen to bring the PC to professional technicians instead of dealing with the city.
They'll check it, adjust the thermal paste on the CPU, and let me know if anything needs replacing or everything is fine.
I'll share their results soon (likely Friday). Thanks for your feedback so far.