Good news for Reboot, BSOD and CTD

,

God News for simmers having

  • sudden reboot PC
  • Blue Screen of Death (BSOD) with WHEA_UNCORRECTABLE_ERROR
  • Crash to Desktop (CTD)

Since Sim Update 5, I wasn’t flying anymore, just trying to find out the reason of different catastrophic events happening without any warning before and any trace after. I was desperate, nothing worked anymore. I followed dozens of recommendations I’ve found here and there, some helped a bit, but no one really. But, today I had a great sense of achievement: After understanding, how power supply of GPU works, I have simply removed the Y-cable delivering power from my PSU to the GPU and replaced it by two single cables. That’s it.

First time after weeks I couldn’t fly at all or maybe only a few minutes, today I had a wonderful 2 hours flight in VR from Innsbruck thru the Alps to Munich with nearly everything set to Ultimate, no Reboot, no BSOD, no CTD, no flickers, no stutters, very smooth, with a happy landing.

Hope this helps some of you!

1 Like

Thats good to hear.

Congrats on solving the issue. Happy flying again.

As an extra tip to yours if you are suffering these CTD etc, try HWinfo. It is good for seeing exactly what is being used. You can also screen dump so others can give advice.

1 Like

Hi @Shojom!

Thanks for the useful topic. I’ve moved it to #self-service:pc-hardware as it’s not a bug report with MSFS, so that others can continue to discuss/add to this topic.

Thanks!

I have my GPU on two single cables right from the build.

Unfortunately I got all the issues anyway, crash to reboot and crash to freeze in the most cases but rarely ctd as well. I can have an 2 hour flight with the 787 one day and hard crash with the TBM 930 the other, with our without addons, with or without OC/UC/UV, with or without Tibetan singing bowls all around my PC …

So this is not the solution here.

I still believe the crashes happen when they happen because they happen. It is the most logical conclusion for me so far.

sehnsedahamses, sorry that my solution did not work for you. But maybe there’s still a chance: Try to use other plugs on your PSU. My experience was, that there is a difference. Unfortunately we never know what’s going on inside the power supply. It may have something to do which plug is connected to which power rail. Probably your two single cables draw power from the same rail as your CPU, what can be a problem, when - under certain conditions in the sim - CPU and GPU have a peak demand at the same time.

2 Likes

There’s one more you can try, sehnsedahamses: If your PSU has a switch to choose between Single Rail Mode and Multiple Rail Mode, try it with Single Rail Mode. In this case there will be no current limit per plug, but every plug can have all the current, at least with my Corsair HX850i. Good luck!

2 Likes

As an addition to this thread, it’s always good practice to connect a separate PCIe power cable from each GPU power slot to your PSU. In other words, use as many cables as you have power ports on your GPU. Daisy chaining these cables, like what OP did, is generally discouraged due to potential heat issues. The increased wattage through a single cable, especially if connected to 3 ports, could cause issues with the wires heating up and potentially melt. This is particularly an issue with the new 3000s because of the high power draw, up to 450W. It’s rare, but there’s been a few posts on Reddit with people claiming their PCIe power cables have started to melt because of daisy-chaining.

1 Like

Thank you for your effort! I have EVGA 850 G+ which is single rail and have connected the GPU by two separate cables on two separate plugs. System draws around 560W max. including monitor and peripherals of what the GPU draws below 350W with some UV/OC 900mV@1860Mhz and 30fps locked. I got up to 99% GPU usage without reaching the power limit and have some % in reserve, the evga 3090 XC Ultra is not the most potent card with its power design. But I was happy to get one at least.

The combined CPU/GPU spikes you note was one of the causes I suspected also. However, I couldn’t check.

I think as long as we are forced to run our GPU’s with brute force for acceptable performance we will have to deal with this crashes.

Upgrading my CPU and GPU, I replaced a smoking 90€ (on sale €60) 750W PSU (got the money back, 750W should be enough, 550 W should be enough) with a 175 € 1000W PSU and you know what, my PC got more responsive (a tad), how does that work … ?

This helped me a lot, sehnsedahamses: OCCT from https://www.ocbase.com/.
It has seperate load tests for CPU and GPU, but also a combined power test, where CPU and GPU are running on 100 % both. This one produced the same BSOD as my FS2020, so I started to investigate about power issues. And now, it’s like in paradise: Today again 2 hours over London with near all Ultimate settings, all problems gone, tested on flat screen (30 FpS locked) and in VR. Best flight simulator experience I ever had.

BTW: I have removed all overclockings, don’t need them anymore. 30 FpS on flat screen are sufficient for me. Don’t know my FPS in VR, but it’s very very smooth.

Very strange, mtjoeng. But strange good news are much better than strange bad news. :wink:

I’m planning to build a new FS-only rig, and yes, it will be a 1000W PSU too.

Thanks, will have a look at it right now. Never heard of it before.