My RTX 3090 Kept Randomly Crashing - Here's How a PSU Upgrade fixed it!

[@beowulf885]

Great video! Explaining how you came to the conclusion that your power supply was no longer sufficient can help a lot of people.

[@SammerJammer]

Glad to see you are back! Looking forward to more videos in the future!

[@johnrucker1912]

Jesus Christ it's Jason Bourne

[@jasonluvisi]

Not surprising at all that a regularly used 13 year old PSU wouldn't be able deliver more than 80% of it's original wattage rating. I'm surprised it still does that well to be honest. I always purchase a PSU rated to deliver 130%-150% of the maximum wattage I think I would ever need... and I replace them at 10 years regardless.

[@Gadtkaz]

for a PSU that old it's likely corsair at the time configured the over power protection limit pretty low. It wasn't until around rdna 1 dropped where manufacturers really started to care about transient spikes knocking out people's systems. Probably says more about how bad ampere's reference power delivery was than anything else as the GN video showed the 3090 spikes pulling twice the normal wattage under load.

[@TechDrivenVideo]

You know that's a really good point! you may have a power supply that can sufficiently power your system, but it's to an older standard that may not work well with newer hardware.

[@kristoffer3000]

That... would explain so much. I have an AX860i and I'm seeing hard pc crashes, straight to pc off with 0 logging in eventviewer. It's a little over 10 years old and undervolting my 3070 has caused it to be a lot less common.

[@abysal311]

A 2019 corsair hx1200i couldn't keep the random crashes away when using a evga 3080 ti ftw3 or a merc 310 rx7900 xtx. Replaced it with new 2024 rm1000x. Solid now. Corsair did RMA and replace the hx1200i, the new 2024 model works fine also.

[@JDD_Tech_MODS]

Had same problem and was most annoying problem ever. The new 850-watt Corsair power supply was the culprit and getting a 1000-watt unit fixed the problem after a month of troubleshooting.

[@JDD_Tech_MODS]

Let me clear up the actual problem was the GPU pulling random transient power spikes while gaming triggered OCP in the 850-watt PSU. No amount of stress testing/GPU tests triggered it and only gameplay did.

[@MATIK000]

how is that so if it still left 100w in, working on full load 640w isn't that 750w one?

[@kristoffer3000]

It also has to power the rest of the pc whilst coping with spikes like that.

[@TechDrivenVideo]

It had around 100watt left but with these spikes no way the power supply can't keep well trying to power the rest of the system. Along with the power supply age after 10+ years of use.

[@lejoshmont2093]

I know you were looking at your total wattage, I am not a power supply expert, but did you check the load per rail to make sure you weren't overloading any one rail? It is what I think off the top of my head that would be pushing the limits of the power supply without exceeding the overall maximum.

[@TechDrivenVideo]

My old power supply had a single rail design so I really shouldn't have been overloading a single rail. It's a good point. I don't have testing equipment to measure each rail and the only thing I could really measure was total power draw.

[@promi4259]

They stil sell this model? Didn't this PSU realesed 10yrs ago?

[@TechDrivenVideo]

I don't think the Ax750 series is sold anymore. It was replaced by Axi series if memory serves me right

[@promi4259]

@TechDrivenVideo i ment the rm1000x

[@R32Danny]

​@@promi4259 this is the newest version of the rm1000x. I think it's the 3rd revision

[@MichelleLowe-f8c]

Power supplies components over time get stressed and capacitors tend to dry out due to heat and power spikes etc. It is best to replace the power supplies after the warranty period!

[@revog7]

a 3080 is typically too much for a 750/850. so would be safe to say a 3090 will crash aswell.

[@crezzlin]

Maybe the thermal paste has gone dry.. I had two 1080ti's that would crash under load after a few minutes.. both had dry thermal paste.. strange thing was both had low temps in MSI afterburner with the crap paste.. Hope you have fixed it with the PSU!

[@TechDrivenVideo]

So far it's been stable. Gpu paste is about 2 years old... But temps are still good so open to trying that if it happens again.

[@Oryzaki0]

I'd highly recommend everyone buy the nicest psu they can afford. In my experience it's the most common point of failure in desktop PCs and the issues they create are not fun to deal with at all. Additionally, unlike most components they almost always retain compatibility with new hardware and thus can last for a really long time.

[@TechDrivenVideo]

I couldn't agree more - a good power supply is definitely worth it ! I just happened to forget to upgrade the power supply along with upgrading all the rest of my components. Like I said in the video

[@TY1979KA]

had the same issue as you did with a corsair rm750x, fixed it with a silverstone vertex gx 1200

[@UpInThePocket8611]

Was having same issues for months..got a new atx 3.0 psu and no more issues

[@johnpaulbacon8320]

Great video. I have always liked to have a higher watagw PSU then what is the so called minimum suggested for the hardware that I currently have and then have that extra for upgrades _ liked what you did with your system.

[@ukjamie96]

That cool and all, but where's Dave? xD

[@TechDrivenVideo]

Back in 2020 I had a chat with Dave about restarting the channel. He chose to not be part of it anymore.

[@mddm40]

Thanks for the video, same for me full shutdown just as Cyberpunk would almost start after i had chosen character then pressed start game blink pc fully shutdown i have Aorus 3080ti master corsair 750x PSU only with Cyberpunk, so yes this helps alot appreciate your time!

[@amigapelit]

when open pc dust cleen to same time

[@liaml1226]

my gtx 1080 wash crashing with my old 500w psu, so i upgraded to 800w and the issue was instantly solved

[@steven4776]

$300 cad?

[@TechDrivenVideo]

Yeah little under that's with tax.