Question GPU fan getting crazy Gigabyte RTX 3070

[Ben_Nachos]

Hello ! One of my fan on my RTX 3070 is litterally going crazy, up to 6000rpm, while the others are fine, does anyone have a clue on what is happening ?
I was playing Ride 4 today, and suddenly it started to go crazy, like for no particular reason except maybe temperature ? I mean, it's 34°C outside, inside is like 25°C, but it never happened to me before in a two years daily use.

PC Specs :
CPU : AMD Ryzen 2700x
Motherboard : Asus ROG X470-F
RAM : 16Gb, 2x8gb, DDR4, G.Skill Trident Z 2400 MHz
GPU: Gigabyte Aorus RTX 3070 8GB (driver version: 512.77 / bios version: 94.4.25.40.83)

 

[Phaaze88]

Both the gpu core and gpu hot spot temperatures are crap. The range is normal though; not unusual to see hot spot 10-20C higher than core.
Memory is not shown - HWINFO should be able to read it. 95C and below is good for that one.


Likely needs a repaste and maybe a good system-wide cleaning. Take it to a shop or a friend who knows how, if you're not comfortable doing the repaste.
Recommended: Noctua NT-H2, Gelid GC Extreme, Prolimatech PK-3, Kingpin KPx, Cooler Master Mastergel Pro V2
There's likely other pastes that work, but I'm not familiar with them.

 

[Ralston18]

PSU: Make, model, wattage, age, condition (original, new, refurbished, used)?

History of heavy use for gaming, video editing, or even bit-mining?

Boot the system and if possible open Task Manager (as Admin) before the fan goes high speed crazy. Watch for what, if anything, changes when the speed increases.

You can do much the same via Resource Monitor. Just do not use both tools at the same time.

Objective being to identify what may be the cause of the speed increases.

One immediate thought is failing thermal paste/pads....

 

[Ben_Nachos]

Both the gpu core and gpu hot spot temperatures are crap. The range is normal though; not unusual to see hot spot 10-20C higher than core.
Memory is not shown - HWINFO should be able to read it. 95C and below is good for that one.


Likely needs a repaste and maybe a good system-wide cleaning. Take it to a shop or a friend who knows how, if you're not comfortable doing the repaste.
Recommended: Noctua NT-H2, Gelid GC Extreme, Prolimatech PK-3, Kingpin KPx, Cooler Master Mastergel Pro V2
There's likely other pastes that work, but I'm not familiar with them.

I've never seen memory temp on HWINFO, I've got the lastest version though.

 

[Ben_Nachos]

PSU: Make, model, wattage, age, condition (original, new, refurbished, used)?

History of heavy use for gaming, video editing, or even bit-mining?

Boot the system and if possible open Task Manager (as Admin) before the fan goes high speed crazy. Watch for what, if anything, changes when the speed increases.

You can do much the same via Resource Monitor. Just do not use both tools at the same time.

Objective being to identify what may be the cause of the speed increases.

One immediate thought is failing thermal paste/pads....

My PSU is a Thermaltake Smart RGB 700W 80+ 230V. All the components on my computer are 4 years old, except for the GPU.

I mostly used it for gaming, I wouldn't call it heavy since I mostly play on the evenings and weekends because work you know !

The fact is that my fans aren't crazy all the time, they just go crazy when a certain temperature is reached, while playing. But, as you can see on the screenshot, is it normal that only GPU#0 gets up to 6000rpm and not GPU#1 ?

So you suggest me to change the thermal paste and thermal pads ?

 

[Phaaze88]

I was sure the devs updated it to allow gpu memory junction temperature readings for RTX 30...

 

[Ben_Nachos]

I was sure the devs updated it to allow gpu memory junction temperature readings for RTX 30...

I mean, maybe I am doing something wrong, but I don't see no options or anything ahah

 

[Ralston18]

Changing thermal paste and thermal pads may be premature.

A four year old PSU is certainly nearing its' designed in EOL (End of Life) and may be starting to falter and fail under load.

Certainly not 24/7 gaming but evenings and weekends can quickly rack up the hours - especially with continual peaking power demands.

Remember that PSUs provide three voltages (3,5, and 12) to various system components. Failure to deliver the necessary voltage and/or watts can result in all sorts of strange behavior.

Take a look in Reliability History and Event Viewer. Either one or both may be capturing error codes, warnings, or even informational events that may correspond with the times the temperature jumps and the fan speeds up.

But also look at the overall pattern: Increasing number of errors and varying errors make the PSU a prime suspect.

Reliability History presents a time line format which can be very revealing.

And, in agreement with @Phaaze88, no harm in cleaning the system and verifying that all connectors, cards, RAM, and jumpers are fully and firmly in place.

Inspect for signs of damage as well.

Lastly: do be sure to back up all important data before doing anything. Backing up is something that should always be being done anyway.

Ensure that that there are at least 2 x backup to locations off of the PC in question. Verify that the backups are recoverable and readable.

PSU or otherwise there is something wrong and a more substantial failure could occur.

 

[DRagor]

I have a simple question: did you verified with your own eyes that in fact one of the fans is rotating considerably faster then the other one?

 

[Ben_Nachos]

Changing thermal paste and thermal pads may be premature.

A four year old PSU is certainly nearing its' designed in EOL (End of Life) and may be starting to falter and fail under load.

Certainly not 24/7 gaming but evenings and weekends can quickly rack up the hours - especially with continual peaking power demands.

Remember that PSUs provide three voltages (3,5, and 12) to various system components. Failure to deliver the necessary voltage and/or watts can result in all sorts of strange behavior.

Take a look in Reliability History and Event Viewer. Either one or both may be capturing error codes, warnings, or even informational events that may correspond with the times the temperature jumps and the fan speeds up.

But also look at the overall pattern: Increasing number of errors and varying errors make the PSU a prime suspect.

Reliability History presents a time line format which can be very revealing.

And, in agreement with @Phaaze88, no harm in cleaning the system and verifying that all connectors, cards, RAM, and jumpers are fully and firmly in place.

Inspect for signs of damage as well.

Lastly: do be sure to back up all important data before doing anything. Backing up is something that should always be being done anyway.

Ensure that that there are at least 2 x backup to locations off of the PC in question. Verify that the backups are recoverable and readable.

PSU or otherwise there is something wrong and a more substantial failure could occur.

Well, I checked the Reliabity History, and there was something called LiveKernelEvent, so I googled it up, some guys said that that might be an error with Nvidia drivers and everything, so I uninstalled everything, and reinstall everything cleanly, and I turned of my pc, checked if everyting was plugged-in correctly, and so far, it seems to have resolved the problem !

Although, I heard you about my PSU, would you advise me to change it soon ? Like, before it causes problems ?

Thank you for your help anyway.

I have a simple question: did you verified with your own eyes that in fact one of the fans is rotating considerably faster then the other one?

And yes I did, only the fan in the middle was overspinning like it wanted to take off !

 

[DSzymborski]

I'm not sure that the PSU is causing the issue, but that PSU with a 3070 sure as hell scares me. That's an unbelievably cheaply made Sirfa of practically ancient design that was unpurchasable until Thermaltake slapped some cheap RGB fans on it and sold it as a budget RGB unit. This PC shouldn't have ever been assembled like this. Even if the PSU isn't the cause, since that PSU ought to be replaced immediately, that's the obvious next step even if it's not the most likely to resolve the original issue.

 

[Ralston18]

Thank you @DSzymborski

= = = =

@Ben_Nachos

All I can add is to be sure that you keep all important data backed up. Which should always be being done anyway....

At least two copies to locations away from the PC/PSU in question. Verify that the data is both recoverable and readable.

 

[Ben_Nachos]

I'm not sure that the PSU is causing the issue, but that PSU with a 3070 sure as hell scares me. That's an unbelievably cheaply made Sirfa of practically ancient design that was unpurchasable until Thermaltake slapped some cheap RGB fans on it and sold it as a budget RGB unit. This PC shouldn't have ever been assembled like this. Even if the PSU isn't the cause, since that PSU ought to be replaced immediately, that's the obvious next step even if it's not the most likely to resolve the original issue.

I totally hear you, but it was 4 years and I was kinda broke, so I had to kinda make choices... have you got any got psu you would suggest me ?

 

[Ben_Nachos]

Thank you @DSzymborski

= = = =

@Ben_Nachos

All I can add is to be sure that you keep all important data backed up. Which should always be being done anyway....

At least two copies to locations away from the PC/PSU in question. Verify that the data is both recoverable and readable.

Heard it ! It already the case about my data and everything !

 

[DSzymborski]

I totally hear you, but it was 4 years and I was kinda broke, so I had to kinda make choices... have you got any got psu you would suggest me ?

True, but you did get a 3070 now. You could have gotten a 3060 or 3060 Ti and a safe PSU!

This is a high-end GPU. Anything from Tier A would do.

Article - [psucultists] PSU tier list rev. 14.8 (Final Update Jul '21)

PSU Tier List 4.0 rev. 14.8 (outdated) Last Update: 12-07-2021 Legend : Gray - EoL/obsolete and/or otherwise not recommended for purchase. Green - small form-factor (gold and blue colors are disregarded due to scarcity of SFX PSUs) Gold - best units in the tier (includes requirements for blue...

forums.tomshardware.com

 

[Ben_Nachos]

True, but you did get a 3070 now. You could have gotten a 3060 or 3060 Ti and a safe PSU!

This is a high-end GPU. Anything from Tier A would do.

Article - [psucultists] PSU tier list rev. 14.8 (Final Update Jul '21)

PSU Tier List 4.0 rev. 14.8 (outdated) Last Update: 12-07-2021 Legend : Gray - EoL/obsolete and/or otherwise not recommended for purchase. Green - small form-factor (gold and blue colors are disregarded due to scarcity of SFX PSUs) Gold - best units in the tier (includes requirements for blue...

forums.tomshardware.com

Yeah, but what I meant is, when I bought the whole computer, I was kinda broke, so I chose the Thermaltake PSU, but didn't thought about changing it when I bought the RTX 3070 !

Thanks, I will change it as soon as I can then