Could be a driver issue or PSU issue at this point, to replicate, just fire up r23 and Furmark (or 3dmark) and run those at the same time, if it shuts down / reboots, we are onto a good winner (as its replaceable).
Then test with an older driver, if still does it, I would very much be looking at the PSU can’t cope.
I also think that there is a very good chance that the PSU could be your problem but obviously I don’t know for sure.
Would it be worth while getting hold of a watt meter (on sale at Amazon for less than £20) to see how much power the PC is drawing at start up? If it’s drawing more power than 460w (i.e. above the spec of your PSU) then maybe this could be a pointer that the PSU could be the source of the problem. If this is the case a continued over load of the PSU could also have led to a faster deterioration of the PSU over a shortish period too especially if it is not of a gold or platinum spec.
If you don’t want to spend any money then I also believe that there is freeware which estimates from your components how much power your PC requires. This will cost you only a bit of your time and might also be an option.
Personally, I would go with the freeware option first and take it from there 
I think if you went to Nvidia directly, and said you were having reboot issues on a 2060 with a 460W PSU, they would point out in the manual you should have at least a 500W.
I have no doubt that you are correct
I have run the Cinebench R23 and all was OK. Then I run the FurMark GPU Stress test and after a few minutes my PC shut down. Then I did another 2 tests with FurMark and all was OK. Does thin mean that my RTX 2060 is faulty?
This is good and bad news, good that its reproducible, bad that its either the PSU or the GPU. If it was the GPU, I would very much expect to see artefacts on the screen before it shut down.
What you could try is downloading MSI Afterburner and lowering the GPU’s power down (there is a slider you can just lower how much power the card can draw), lower it by 50 watts (so probably 25% on the slider if the card is a 200w TDP) and then try that, yes you will lose performance, but as a test, it will tell you at least if its PSU or GPU related (if it shuts down then chances are its the PSU).
Another simple test would be to run r23 and Furmark at the same time, if it crashes a lot quicker than just the Furmark running its very much pointing towards the PSU.
PSUs, infact all hardware degrades over time and two years for a laptop is not uncommon but you can also help things by keeping fans and heatsinks free of dust and fluff, using a cooling mat etc. Also I can’t remember which Alienware model but Dell was caught secretly undervolting just to save them upgrading it’s power supply and or cooling.
OK Thanks for your suggestions. I will try that. One thing though that still puzzles me is why (if the PC doesn’t reboot when switched on) it never reboots after that even with the MSFS on I am on a flight now for about 45 minutes and all is OK.
Could be down to transient loads, or things have warmed up (I had a GPU back in the day that would always fail on a cold boot but once warmed it was fine — bad solder connections).
Its 90% a PSU issue at this point.
I think Nvidia tends to overestimate PSU requirements to be on the safe side. To account for overclocking and all that good stuff. Since we’re talking about Dell here, it’s probably an OEM graphics card also.
I’m suddenly experiencing same issues after last Nvidia drivers update, my pc randomly reboots and I can’t find the issue. Will try to revert driver version after work and see if it helps. It even rebooted while I was checking the system temperature (32c) and GPU idle.
RTX3070 Suprime
i9 10900K
850W
32GB RAM
Assuming that’s the gpu lanes that does not look good so check addresses. However (at a guess) if your main (first) M.2 slot is only occupied by a non NVMe SSD it could even be normal. Maybe you should speak to your vendor.
Hi Can you please say it in more simple terms that I can understsnd it?
“Assuming that’s the gpu lanes that does not look good so check addresses.” ???
“your main (first) M.2 slot is only occupied by a non NVMe SSD it could even be normal” I have only one slot for NVMe .
You need to know if your gpu is using bus 1. And some boards have more than one M.2. slot but not all. Your board seems to support PCIe.4 although I can’t be certain of NVMe so read the manual but if it does and your drive is just PCIe.3 NAND I might expect a similar warning.
I recommend checking for bios, motherboard (inc. chipset) and processor updates and firmware for your SSD. There is also a tiny chance that your mainboard’s coin sized battery needs replacing.
I’ll just add that I don’t trust Dell since they were caught secretly throttling high end gpu’s just to fit their power curve.
Did you DDU before reinstalling Nvidia? If that doesn’t help it’s most likely a hardware problem although I would tend to suspect the MB’s chipset rather than the gpu itself. The exception would be unless the card is worn out from continuous bitcoin mining.
Looking on the web, its not good. Looks like others have had this issue and there is nothing you can do to sort this, something is broken (GPU or Motherboard).
If its a tower, just open the case up and reseat everything, the GPU, memory and make sure PSU cables are all snug (reseat them)
If that fails, Id contact Dell and get it looked at (this is not a driver issue at this point as you are not in windows).
good point although there are basic ROM drivers attached to the bios chip.
Well I am going to take my PC to the DELL agents here as it is still under warranty…
If they format it and reinstall windows, will I be able to redownload MSFS?
Yes … so long as you are signed into the same MS account you should.
The only problem I can foresee is if your PC is automatically given a different name but even then it shouldn’t matter (don’t forget to delete the old one in your MS account settings). For Steam you may have to do backflips and make three Hail Marys .. who really knows?
