koolisdon2006
Ok thanks. So if it runs in debug mode it most likely will be the power supply?

Sajin
Swapping the psu out with a brand new one will help you narrow down the issue. 850w is plenty. This problem could be coming from the card itself, overheating hardware, or even an unstable oc.

bret0213
This is an old thread, but I have the same issue with my 2080ti black edition.
 
My setup is 7700k, ASUS z270i, 2080ti Blk, 32gb corsair 3200mhz, Nvme 970 pro, Nvme WD black, Silverstone sfx700w Lian-li Riser.
 
 
The only solution that I found to fix the problem is to decrease the power limit to 90% and set max frames to display at 144hz. The GPU runs at 1790-1890 during gaming sessions with 90% power limit. Same clock at 100% PL before crashing, but doesn't happen has frequent has overclock. Increasing the voltage to 100% and power limit to 112% with +100mhz and +600 memory shows stability running at 2050mhz and can run benchmarks fine. Has soon has I enter a live video game though it will reboot randomly, last longer when the frames are restricted but still happens. Only get kernal 41 (I believe, whatever kernel is lost of power), nothing else.
 
I have RMA the PSU and then replaced it with a Silverstone sfx-l 800w, but had the same problem.
 
These are all the other trouble shooting have done with no success: tried different wall sockets, used two PCIE PSU cables instead of a splitter. tried different gauge psu cables. Reinstalled windows. clean install drivers. underclock cpu. increase voltage to cpu. decrease voltage to cpu.tried a couple different risers cables. ran multiple memory test, swapped memory slots, swapped memory for 16gb running at 2133mhz, removed nvme and used a sata drive, swapped MB, double check temperatures and increase fan curve.
 
I am waiting for a corsair 750w psu to try the different rail. My power drawl seems to be good for the size of PSU (cards max watts is 234-250w non overclock, 284w with some spikes to 300w at overclock, the rest is around 234w) and I am limited in a SFF build I have. The card seems like a silicon lottery because of the boost speed its getting with a power restriction, I really don't want to go through the RMA process, but don't know what else to do. Can anyone think of anything else? The only other thing I can think of before RMA is taking it a part and putting thermal pads on the vrms and memory and replacing paste, but won't that void my warranty? Sensors don't get above 71C overclocked but maybe something else is heating up and causing a voltage issue... idk

R3P01

bret0213
Well the PSU just came in shortly after my last post. These are initial results but it looks like it might have been the Silverstone rail. I don't think it was anyone particular PSU from Silverstone due to having used 3 of them. 2 at sfx-700g and one sfx-l-800t. Again this is preliminary, have only tested for about an hour but usually reboots would happen within first 5 minutes.

Test it for a few days. EVGA told me my 1660ti issue was the 12v rail as well. I replaced my PSU and have now RMA 2 cards, 3-4 days later my new RMA card started losing video signal again and crapping out. I have another RMA card OTW now to be delivered Tuesday.

bret0213
 
The Corsair PSU solved it all. My conclusion is something is wrong with Silverstone's rail, considering the three different Silverstone PSUs that I tried. Surprising, Silverstone is known for their quality and I don't think anything was wrong with any of the PSU per say, just can't handle these power hungry cards that NVIDIA is coming out with.