Question Computer keeps crashing when playing VR games ?

[brandonmacleod21]

This happens most when I play Half Life: Alyx where my computer would make a very loud clicking/popping sound from my headphones and all my displays would turn off. My computer would remain turned on but it needs to be turned off and manually restarted. It also appears that my motherboard has a red light on it when this happens. Originally, my RAM was at 64gb running at a speed of 6400MHz. At this speed, Half Life: Alyx would crash immediately. I then reduced the RAM speed to 4800MHz and I was able to play the game for a little bit longer but the computer still eventually crashed. I then reduced this speed again to 3400MHz and it looked like everything was working.

However, after a little while, my computer crashed again while playing a VR game. After I restarted the computer, I then proceeded to go into SteamVR and attempted the room setup (because I recently moved my base stations) and every single time, my computer crashed. I then added another 64gb of RAM and that seemed to have worked.

However, I recently started up Half Life: Alyx and the computer crashed when I was just on the main menu. I then took out all the RAM and put in some old RAM I had which was 4 16gb ram sticks and that seemed to have fixed the issue. I played the game for several minutes and then exited it and turned off my base stations. SteamVR was still running while I started up Google Chrome and then it crashed again while the browser was starting up. The only other time it crashed when I wasn't running a VR game was when I was editing something in Photoshop (the only thing was that my VR headset was still plugged in but that was it).

It seems as if changing or adding RAM (or lowering the RAM speed) fixes the issue but only for a few days and then it keeps crashing again. I never used to have this problem until I got my VR headset. Does anyone know what the problem is? Here are my other specs.

GPU: RTX 4090

CPU: Ryzen 9 7950X

OS: Windows 11

 

[Ralston18]

Full specs please: Motherboard, RAM, Disk drive(s), PSU, attached peripherals.

One good starting point is to look in Reliabililty Histor/Monitor for error codes, warnings, and even informational events being captured just before or at the time of the crashes.

Increasing numbers of errors and varying errors make the PSU suspect.

However, more needs to be known.

 

[brandonmacleod21]

Full specs please: Motherboard, RAM, Disk drive(s), PSU, attached peripherals.

One good starting point is to look in Reliabililty Histor/Monitor for error codes, warnings, and even informational events being captured just before or at the time of the crashes.

Increasing numbers of errors and varying errors make the PSU suspect.

However, more needs to be known.

I have 3 solid state drives where two are Sata drives and the main C drive is an NVMe. I also have two more external drives plugged in via USB. Attached to the computer are a pair of headphones, a mouse and keyboard, 3 monitors (24 inch, 42 inch, and 32 inch), and the VR headset. The motherboard is a B650 Aorus Elite AX with 64gb of ddr5 RAM at the moment (although I did upgrade it to 128gb and that worked for a little while until it started to crash again). My PSU is an AGT1000 which is a 1000 Watt power supply.

I looked at Reliability Monitor but it doesn't seem to be very helpful since it just says that Windows was not shut down properly and the technical details says "The previous system shutdown was unexpected." This was most likely due to me shutting it down manually since all my monitors go black (and I can't get them to display anything no matter what I do) and I have to turn off the power strip to restart it (since holding down the power button doesn't work at that point).

 

[Ralston18]

My immediate thought is power problems.

The PSU is simply not able to keep up with power demands and especially sudden peaks in power demands.

I am not familar with the AGT1000 PSU so will need to defer to those who are...

What can and does happen with unexpected shutdowns is that system and app files can be corrupted. Windows must go through a specific series of actions during the normal shutdown process as a means to prepare for the next startup.

If those actions are incomplete or otherwise disrupted (loss of power) then problems occur. And all gets worse.

Likely increasing numbers of errors and varying errors.

Look in Event Viewer - not as user friendly as Reliability History/Monitor and requires more time and effort to navigate and understand.

To help:

How To - How to use Windows 10 Event Viewer | Tom's Hardware Forum (tomshardware.com)

Objective being to discover errors that occurred before you were forced into a manual power off shutdown.

= = = =

You can use the built in Windows troubleshooters along with "dism" and "sfc /scannow" to find and fix such problems.

FYI:

https://www.windowscentral.com/how-use-dism-command-line-utility-repair-windows-10-image

https://www.lifewire.com/how-to-use-sfc-scannow-to-repair-windows-system-files-2626161

The first thing you should do (as should always be being done) is to ensure that you have all important data backed up. And proven recoverable.

Then, if at all possible, swap in another known working (no problems) PSU. Be sure to use only the cables that come with the swapped in PSU.

One other note: Take a careful look at how all your devices are connected. Power, network, audio, video, peripherals cables and ports.

Look for some sort of loop: Device A connected to Device B connected to Device C connected to Device D connected to Device A again. Any path of connections.

Especially if using surge protectors/power strips. Most are low end and can be or become problematic.

 

[COLGeek]

I think @Ralston18 is on target. This seems like a possible PSU issue. The 4090 is very demanding and the AGT1000 seems to be a low-cost, mediocre PSU.

This is almost certainly a heat and/or power issue, as described. If overclocking anything, stop and revert to stock speeds (set BIOS to default settings). You want a stable baseline, if possible, before tweaking things.

Another thing to do is to disconnect the additional displays. More displays means more pixels to manipulate and that equals more heat/power. Once again, intent is to get to a stable baseline and proceed from there.

 

[brandonmacleod21]

I think @Ralston18 is on target. This seems like a possible PSU issue. The 4090 is very demanding and the AGT1000 seems to be a low-cost, mediocre PSU.

This is almost certainly a heat and/or power issue, as described. If overclocking anything, stop and revert to stock speeds (set BIOS to default settings). You want a stable baseline, if possible, before tweaking things.

Another thing to do is to disconnect the additional displays. More displays means more pixels to manipulate and that equals more heat/power. Once again, intent is to get to a stable baseline and proceed from there.

I'm looking into things right now but the only other PSU that I can put in is a 10 year old 450W power supply. For a 4090, if I were to get a new PSU, would you recommend getting one that is more powerful or is there one that is known to be more stable?

 

[COLGeek]

I'm looking into things right now but the only other PSU that I can put in is a 10 year old 450W power supply. For a 4090, if I were to get a new PSU, would you recommend getting one that is more powerful or is there one that is known to be more stable?

Try the other things mentioned as well.

Wattage should be okay. Quality is more important. What is your budget?

In the meantime, for your consideration:

https://www.corsair.com/us/en/p/psu...m-atx-1000-watt-pc-power-supply-cp-9020259-na

 

[brandonmacleod21]

Try the other things mentioned as well.

Wattage should be okay. Quality is more important. What is your budget?

In the meantime, for your consideration:

https://www.corsair.com/us/en/p/psu...m-atx-1000-watt-pc-power-supply-cp-9020259-na

I also forgot to mention this but my desktop is currently hooked up to a UPS. The UPS is 1000 Watts.

 

[COLGeek]

I also forgot to mention this but my desktop is currently hooked up to a UPS. The UPS is 1000 Watts.

Okay. Try this as well, play system without using UPS. One of our moderators had a similar issue a while back. Endless troubleshooting. Nothing helped.

On a whim, he removed the UPS from the equation. Voila!

UPS was underpowering system under full load.

Replaced with a stronger unit.

 

[brandonmacleod21]

Okay. Try this as well, play system without using UPS. One of our moderators had a similar issue a while back. Endless troubleshooting. Nothing helped.

On a whim, he removed the UPS from the equation. Voila!

UPS was underpowering system under full load.

Replaced with a stronger unit.

OK, I removed the UPS and the power strip from the equation and I disconnected my other two monitors as well as lowered the resolution of my main monitor and it still crashed. I then reset the BIOS settings and then re-enabled CSM support (and disabled secure boot) because I couldn't login unless it was enabled. It should be noted that my CPU used to be undervolted because it was constantly overheating and I had lowered my RAM speed from 4800MHz to 3400MHz. With the BIOS reset, my RAM is back up to 4800MHz and the CPU clock speed and voltage is back to its default values.

I plugged in my controller, turned on my base stations, and turned on the controllers. Shortly after SteamVR started up, my computer crashed again. It almost seems worse now since I wasn't even running a game and it still crashed. Since my computer was still technically running after it crashed, I noticed that there was a red LED on my motherboard that wasn't lit up before so I took a picture of it (ignore the glare, that was from the camera flash). Do you know what this red LED means?

 

[COLGeek]

What is the make/model of your motherboard? With that info, we can look up your motherboard manual and likely determine what the LED is indicating.

 

[brandonmacleod21]

What is the make/model of your motherboard? With that info, we can look up your motherboard manual and likely determine what the LED is indicating.

It's a Gigabyte B650 Aorus Elite AX AM5.
EDIT: Based on the manual for this motherboard, it looks like the DRAM LED is lit up.

 

[COLGeek]

See page 19 of the manual.

https://download.gigabyte.com/FileL...1202_e.pdf?v=10b7e5f3fa0c69cbc4eded7b7279eeb7

Seems to be the DRAM (memory) LED.

 

[brandonmacleod21]

See page 19 of the manual.

https://download.gigabyte.com/FileL...1202_e.pdf?v=10b7e5f3fa0c69cbc4eded7b7279eeb7

Seems to be the DRAM (memory) LED.

That would imply that it's a memory issue but I swapped out the memory several times and it didn't fix anything.

 

[COLGeek]

That would imply that it's a memory issue but I swapped out the memory several times and it didn't fix anything.

Is your motherboard BIOS currently configured for default settings for the memory?

Also, at any point while troubleshooting, have you removed the CPU and inspected the socket for damaged/missing pins?

 

[brandonmacleod21]

Is your motherboard BIOS currently configured for default settings for the memory?

Also, at any point while troubleshooting, have you removed the CPU and inspected the socket for damaged/missing pins?

I just replaced the PSU and so far, it seems to be working. I'll give it a few more days just to be sure.