Question Random and persistent system crashes/freezes on brand new setup

[dancastro88_]

Hi all,

This is a new build from June that worked perfectly throughout the month until Tuesday of last week. Since then I I have had daily crashes under low/no load and attempted multiple fixes and forums. Hoping you all can shed some light.

Issues encountered:
- Screen freeze followed by blacking out and powering off.
- Screen freeze and required forced shutdown
- Screen goes to sleep (while all power off and sleep settings are "never") and then comes back on
- When rebooting from a crash, wont get past the MB screen.
- System freezing while browsing under very low load / Freezing when opening start menu / Freezing and crash when just opening excel.
- When installing AMD bluetooth driver update I lost ethernet connection.
- While installing windows 11 from USB, system refusing to restart and screen flickered green.
- Extreme lag in windows log followed by desktop (usually ended up in system crashing)


System Specs:
Case: Lian Li A4 H20
Board: B650i Aorus Ultra
BIOS: BIOS Version/Date American Megatrends International, LLC. F30, 22/05/2024
VGA: MSI Ventus 3x OC 4070TI Super 16gb
Riser Cable PCIE Gen 4 - included in case
PSU: Corsair SF750
CPU: AMD Ryzen 7800x3d
MEM: G.SKILL Flare X5 Series (AMD EXPO) DDR5 RAM 32GB (2x16GB) 6000MT/s CL30-38-38-96 1.35V (F5-6000J3038F16GX2-FX5)
HDD/SSD: 1x Ssd Samsung 980 Pro 1tb M.2 Pci-e Gen4 Nvme - 7000mbs
COOLER: Water Cooler Cooler Master Masterliquid 240 Atmos Argb 240mm
Keyboard: At the moment a Logitech G pro TKL
Mouse: Logitech Pro Wireless
OC: None
OS: Windows 11 64bit Home version 23H2
Display: ASUS TUF 31.5", 144Hz, 2K QHD, 1ms, DisplayPort e HDMI, FreeSync, HDR
- How display is connected to the GPU?: DP

What I've tried and did not help:
- Made sure I am on latest BIOS (there is 1 version recently from July I still need to test)
- 2x windows installs, followed by AMD chipset updates and clean nvidia drivers. Done from a newly created bootable USB install.
- DDU and clean install of latest nvidia drivers
- DDU and install of 2 previous nvidia drivers
- Changed power outlet
- Reset GPU in case and repositioned cables. Checked connection in PSU and all power connections in MB. Checked PCIE connections to 12vhr split and repositioned slightly in case.
- Currently doing a memtest86 on each RAM. Card 1 is 90% done and has zero errors.

What I've tried and seems to work
- My previous system had a GTX 1080TI, when installing in my current build the system seems to work. I had 1 crash only a few days ago but it was not clear why. I have had this system running for a while and had no further issues..
- Installing 4070ti super in my older system - So far zero crashes at all. In this build the GPU sites on the MB.

Sharing a link with the most recent crashes on the system I described. Before the crashes I ran 3d mark tests that worked fine only for the system to freeze and crash randomnly later in the morning.
Link: New System with 4070

I have had the 4070 in my old system all day today with zero errors. Similarly, had the 1080ti in my new system with no issues for most of the day, I asked a question on this in the microsoft forums and the answer was this was a driver issue followed by a likely bad card (but the card works fine in a different setup).

On MSI the folks pointed to a possible issue with the Riser cable, but the pc was working fine for 1 month before any issues at all and the 1080ti is on the same riser.

I am hoping you all can help me before I give up and have to take the pc to repair shop.

Thank you!

 

[Lutfij]

Welcome to the forums, newcomer!

BIOS: BIOS Version/Date American Megatrends International, LLC. F30, 22/05/2024
You have one more BIOS update pending.

2x windows installs, followed by AMD chipset updates and clean nvidia drivers. Done from a newly created bootable USB install.
Did you install the OS in offline mode? Manually installing all relevant drivers sourced from their respective support sites?

PSU: Corsair SF750
Is this unit brand new?

I asked a question on this in the microsoft forums and the answer was this was a driver issue followed by a likely bad card (but the card works fine in a different setup).
You can rule out the GPU as your root cause by dropping it into a system that has more power from the PSU at the entire system's disposal.

On MSI the folks pointed to a possible issue with the Riser cable, but the pc was working fine for 1 month before any issues at all and the 1080ti is on the same riser.
If you want to rule out the riser cable, breadboard the system without the riser cable and see if the issue persists.

 

[dancastro88_]

So today I finished memtest86 on both sticks of ram and no issues found. After I inserted both noticed the PC is starting up very inconsistently. Sometimes I can get into bios, sometimes straight to windows, on restart from windows often the PC will just not restart and I can see the keyboard light up and screen go on and off until I have to force shutdown.

Was a struggle to get into the BIOS and update to the latest version which went weirdly. The pc restarted after I started the process and screen just kept going on and off, I tried switching from HDMI to DP but never got back to the bios update screen. Eventually made it directly into windows and from msinfo32 I can see the updated version is there. Tried a 3d mark test and scores we worse than before touching RAM or updating BIOS.

PSU is brand new, all the components are new.

I installed windows 11, needed internet connection. But once I was in windows I would allow it to update fully, then update nvidia drivers. AMD chipset and others I would update directly from gigabyte website for my mb version rev.1.

GPU I think I have ruled out as I have installed in my old system for 2 days straight now and no issues. All the inconsistency in start up, odd power on and off is concentrated on the new PC.

Riser cable is on the list, I noticed something that looks like damage to the cable (not sure though). I will be taking the system in for repairs as I have given up.


https://photos.app.goo.gl/bToDS3DYTaXy3t4r7

 

[helper800]

So today I finished memtest86 on both sticks of ram and no issues found. After I inserted both noticed the PC is starting up very inconsistently. Sometimes I can get into bios, sometimes straight to windows, on restart from windows often the PC will just not restart and I can see the keyboard light up and screen go on and off until I have to force shutdown.

Was a struggle to get into the BIOS and update to the latest version which went weirdly. The pc restarted after I started the process and screen just kept going on and off, I tried switching from HDMI to DP but never got back to the bios update screen. Eventually made it directly into windows and from msinfo32 I can see the updated version is there. Tried a 3d mark test and scores we worse than before touching RAM or updating BIOS.

PSU is brand new, all the components are new.

I installed windows 11, needed internet connection. But once I was in windows I would allow it to update fully, then update nvidia drivers. AMD chipset and others I would update directly from gigabyte website for my mb version rev.1.

GPU I think I have ruled out as I have installed in my old system for 2 days straight now and no issues. All the inconsistency in start up, odd power on and off is concentrated on the new PC.

Riser cable is on the list, I noticed something that looks like damage to the cable (not sure though). I will be taking the system in for repairs as I have given up.


https://photos.app.goo.gl/bToDS3DYTaXy3t4r7

I would highly suspect the riser cable. This is because the 1080 ti is a PCIe 3.0 card and the 4070 ti super is a PCIe 4.0 card. A lot of risers can have problems with PCIe 4.0 cards and its much higher data throughput. A bad PCIe 4.0 riser cable can play nice with lower bandwidth cards like a 1080 ti and then cause all sorts of issues like this with a full bandwidth card. This is a known issue with lower quality controlled riser cables.

 

[dancastro88_]

I would highly suspect the riser cable. This is because the 1080 ti is a PCIe 3.0 card and the 4070 ti super is a PCIe 4.0 card. A lot of risers can have problems with PCIe 4.0 cards and its much higher data throughput. A bad PCIe 4.0 riser cable can play nice with lower bandwidth cards like a 1080 ti and then cause all sorts of issues like this with a full bandwidth card. This is a known issue with lower quality controlled riser cables.

Is it typical that the riser cable would start malfunctioning after 1 month of use or is this usually an issue that is apparent immediately?

 

[helper800]

Is it typical that the riser cable would start malfunctioning after 1 month of use or is this usually an issue that is apparent immediately?

It depends on exactly what is wrong, if anything, with th riser cable. You said you noticed damage of some sort on the riser cable? Expand on that. It could be that the cable got kinked/pinched and the internal copper traces/wires are damaged enough to cause such issues.

 

[dancastro88_]

It depends on exactly what is wrong, if anything, with th riser cable. You said you noticed damage of some sort on the riser cable? Expand on that. It could be that the cable got kinked/pinched and the internal copper traces/wires are damaged enough to cause such issues.

shared an image before: https://photos.app.goo.gl/bToDS3DYTaXy3t4r7

It seems like some of the copper is exposed, but not sure if that is any real damage or if it was there when I bought the case. Once I had built this early June I hadn't opened the case until a few days after the first (apparent) nvidia driver error.

That's what's driving me mad. I dont get what triggered this all.