[SOLVED] GPU fan keeps revving for no reason

[gm matthew]

I recently bought and installed a Sapphire Dual-X Radeon R9 270X for my PC. While I'm pleased with the improved graphics performance over my old card, there's one thing bugging me: occasionally, the GPU fan will spin up to maximum speed for seemingly no reason at all, causing lots of undesired noise.

These bursts of speed seem to happen at complete random, even with the GPU at room temperature. I've tried tinkering with both AMD OverDrive and Sapphire TRIXX, and neither of them can stop the fans from randomly spinning at 100%, no matter what settings I use.

Sapphire TRIXX reports the current fan speed as a percentage, but presumably this is just what the fan is being told to do as it can't detect the random bursts of speed. For example, I set the fan speed to 20%, it says it's doing 20%, but the loud whirring and HWMonitor tell me otherwise.

It's left me completely stumped; any ideas? (Yes, my drivers are up-to-date.)

EDIT: It turns out that the revving was caused by a loose cable on the card; I simply plugged it in more firmly and the fan has been working beautifully ever since.

 

[Grayfox]

Updated drivers?
Check google for any other users with the same issue and see if they have a fix.

 

[gm matthew]

Updated drivers?
Check google for any other users with the same issue and see if they have a fix.

Checked Google, no luck. It seems most other users with this problem simply have overheating GPUs.

 

[Coxis]

even with the GPU at room temperature.

Do you live in a tropical country or something? I can't see a modern GPU idling at less than 30C!

In all seriousness though, if I were you I'd stop looking for workarounds like manually controlling fan speed and exchange it for a new one, especially if you just bought it. If you are monitoring its temperature and the fan goes crazy even when it's not hot / under heavy load then there's probably something wrong with the card.

Having said that, if you want to do further testing before returning it I'd recommend MSI afterburner.

http://event.msi.com/vga/afterburner/overview.htm

 

[terminator363]

Do you live in a tropical country or something? I can't see a modern GPU idling at less than 30C!

In all seriousness though, if I were you I'd stop looking for workarounds like manually controlling fan speed and exchange it for a new one, especially if you just bought it. If you are monitoring its temperature and the fan goes crazy even when it's not hot / under heavy load then there's probably something wrong with the card.

Having said that, if you want to do further testing before returning it I'd recommend MSI afterburner.

http://event.msi.com/vga/afterburner/overview.htm

MSI Afterburner

 

[TenEightyOne]

Sapphire TRIXX reports the current fan speed as a percentage...

Which version of Trixx are you running? Your fan speed also sounds very low, how does the card perform at about 60%?

 

[gm matthew]

Do you live in a tropical country or something? I can't see a modern GPU idling at less than 30C!

In all seriousness though, if I were you I'd stop looking for workarounds like manually controlling fan speed and exchange it for a new one, especially if you just bought it. If you are monitoring its temperature and the fan goes crazy even when it's not hot / under heavy load then there's probably something wrong with the card.

Having said that, if you want to do further testing before returning it I'd recommend MSI afterburner.

Okay, slightly above room temperature then. In any event, the GPU isn't overheating.

I ordered the card just over 30 days ago and chose not to get a warranty, so no returning it now. Not sure if MSI Afterburner will help much (I chose TRIXX since I have a Sapphire card), but I might give it a try just in case.

Which version of Trixx are you running? Your fan speed also sounds very low, how does the card perform at about 60%?

I'm using v4.8.6. On the automatic setting, the fan usually spins at 20%, which is enough to keep the GPU cool when it's not really doing anything. (The actual "20%" speed is 1100 RPM or so, compared to a 100% speed of 3400 RPM.) Choosing a fixed speed, whether it's 20% or 60%, doesn't stop the fan from revving to 100% every now and again.

Thanks for the responses guys 👍

 

[TenEightyOne]

Okay, slightly above room temperature then. In any event, the GPU isn't overheating.

I ordered the card just over 30 days ago and chose not to get a warranty, so no returning it now. Not sure if MSI Afterburner will help much (I chose TRIXX since I have a Sapphire card), but I might give it a try just in case.

Definitely give MSI a try if you haven't!

The warranty is irrelevant if you purchased it to a UK address, in addition to CCR (which supersedes Distance Selling Regs) you have the normal two year warranty conferred by the Sale of Goods acts. Actually up to six years... manufacturers don't like you to know that

Check the card out with MSI then write an email to the supplier if you have no joy

 

[JackBlack90]

Hello 2014!! I wanted to post because I have been unable to find anyone else with the same problem until now. I basically have the same problem but with an xfx 390x DD.

I RMAed it to the retailer and they said they left it idle for a day and it didn't rev up. However, they left it playing youtube all day and did not put the card under any load (which may be important). My problem is also apparently NOT related to the current temperature of the card.

I put a 2 year old MSI 270 in the machine and it has flashing textures and artifacts but no apparent revving.

Thanks to your post and in particular, your edit, I am starting to suspect that I have some kind of power issue. I know that the other card didn't rev but maybe the artifacts were the same thing but in different form idk. Ive taken the card out about 3 times now, so I am pretty sure the cables are snugly placed in the socket. I wonder if dust could cause this issue? A faulty power cable? Or maybe a faulty PSU?

Just to clarify as well, I assume the loose cable you refer to was the extra power lead needed to power your GPU?

 

[Grayfox]

Hello 2014!! I wanted to post because I have been unable to find anyone else with the same problem until now. I basically have the same problem but with an xfx 390x DD.

I RMAed it to the retailer and they said they left it idle for a day and it didn't rev up. However, they left it playing youtube all day and did not put the card under any load (which may be important). My problem is also apparently NOT related to the current temperature of the card.

I put a 2 year old MSI 270 in the machine and it has flashing textures and artifacts but no apparent revving.

Thanks to your post and in particular, your edit, I am starting to suspect that I have some kind of power issue. I know that the other card didn't rev but maybe the artifacts were the same thing but in different form idk. Ive taken the card out about 3 times now, so I am pretty sure the cables are snugly placed in the socket. I wonder if dust could cause this issue? A faulty power cable? Or maybe a faulty PSU?

Just to clarify as well, I assume the loose cable you refer to was the extra power lead needed to power your GPU?

Revving could be just driver related issues so update those but then again AMD doesn't make the best drivers for their GPUs.

Your old R9 270 has an issue with VRAM if it is getting flashing textures and artifacts

Install MSI Afterburner and use a custom fan profile so the fan sticks to a set speed for a given temperature range.

 

[Nexus]

It might be a power supply problem, a friend of mine was also having fan problems with his brand new 1070. I don't think it's a driver problem, he tried downgrading drivers and even did a reinstall of Windows, but the problem was still there.

Revving could be just driver related issues so update those but then again AMD doesn't make the best drivers for their GPUs.

I thought this meme died a few years ago.

 

[JackBlack90]

Hi guys,

Many thanks for the replies and suggestions.

I have since installed the graphics card again for the third or fourth time and it seems to be working a bit better this time, touch wood. It revved very briefly once or twice.

Originally, I installed different drivers and that had no effect. On the latest installation I installed the newly released drivers. I also completley disconnected and reconnected one VGA cable to the PSU.

I installed MSI afterburner and set a custom fan profile etc. This did not override the problem. It would (bizarrely) change the fan speed only when there was no problem.

I still think that the issue isn't resolved but at the moment it is not nearly as bad as it was. I will probably try reconnecting all power leads, cleaning a bit more dust off, multimeter or RMA the PSU. I was planning to reinstall windows anyway so i'll do that as well.

 

[Grayfox]

The video card may need a bios update but do NOT do this unless you know what you're doing and are aware of the risks involved.
But you did say the card is the "390" I can only assume this is the AMD R9 390 which has also been known as the Room Heater as it was made to run at 80°C under load from what I have read.

@TwinTurbo LM It is a fact that AMD drivers were bad esp for the R9 390, but that is not too say nVidia's drivers are good either.

 

[JackBlack90]

The video card may need a bios update but do NOT do this unless you know what you're doing and are aware of the risks involved.
But you did say the card is the "390" I can only assume this is the AMD R9 390 which has also been known as the Room Heater as it was made to run at 80°C under load from what I have read.

@TwinTurbo LM It is a fact that AMD drivers were bad esp for the R9 390, but that is not too say nVidia's drivers are good either.

There is a chance that I don't know what I am doing with a bios update so I will probably just leave that until I have tried everything else.

Tbh I think 390x temps are pretty normal for a card of its generation and tier. The 980 often runs at 80. The 290x ran at 95 degrees or something crazy though. 390x is a revised version of the 290x with better cooling and about 10% better performance coming from higher clock rates and memory frequency (I think).

I have heard that AMD drivers suck. I'll probably get an nVidia next time although it depends on the products being offered.

It revved this morning for like 5 seconds. That is the second time it has done this since I put it back in a week ago. That is not really a big deal compared to before but still would be nice to get to the bottom of this.

 

[Grayfox]

There is a chance that I don't know what I am doing with a bios update so I will probably just leave that until I have tried everything else.

Tbh I think 390x temps are pretty normal for a card of its generation and tier. The 980 often runs at 80. The 290x ran at 95 degrees or something crazy though. 390x is a revised version of the 290x with better cooling and about 10% better performance coming from higher clock rates and memory frequency (I think).

I have heard that AMD drivers suck. I'll probably get an nVidia next time although it depends on the products being offered.

It revved this morning for like 5 seconds. That is the second time it has done this since I put it back in a week ago. That is not really a big deal compared to before but still would be nice to get to the bottom of this.

90°C is high for a GPU IMO, I have heard that transistors start to fail after 70°C

 

[BobK]

The 980 often runs at 80. The 290x ran at 95 degrees or something crazy though. 390x is a revised version of the 290x with better cooling and about 10% better performance coming from higher clock rates and memory frequency (I think).

The maximum recommended operating temperature for the "typical CMOS part" is 85°C. So at 90 or 95 you can expect premature failure of the card.

90°C is high for a GPU IMO, I have heard that transistors start to fail after 70°C

Sounds about right.

 

[JackBlack90]

The maximum recommended operating temperature for the "typical CMOS part" is 85°C. So at 90 or 95 you can expect premature failure of the card.



Sounds about right.

Thanks for reply. I made mistake, 290x typically 70s/80s. Only sometimes 90s. (I have 390x @ 80 °C)

Many cards run @ 70s/80s so surprised to learn that start to fail > 70.

 

[Grayfox]

Many cards run @ 70s/80s so surprised to learn that start to fail > 70.

Is that many AMD cards or many cards in general?.


My old EVGA GTX680SC never peaked 70°C and it had the reference blower cooler on it.

 

[Nexus]

Is that many AMD cards or many cards in general?.


My old EVGA GTX680SC never peaked 70°C and it had the reference blower cooler on it.

I assumed that it applied to all 28nm cards. According to this review, the load temperature of the 680 was closer to 80°C. Maybe your card was being throttled?
Also, the GTX 480 had a load temperature of 90°C, I don't recall that being a problem.

 

[Grayfox]

I assumed that it applied to all 28nm cards. According to this review, the load temperature of the 680 was closer to 80°C. Maybe your card was being throttled?
Also, the GTX 480 had a load temperature of 90°C, I don't recall that being a problem.

Nope.

But the CPU was water cooled so there would have been less heat in the case.

 

[Goshan111]

I experienced the same trouble with Asus 1060 OC 3gb - intermittent fans revving when setting Afterburner fan slider to idle (or auto). (Actually, I tried using Afterburner to solve another issue - one of the two fans sometimes would work at about 20% constantly, while the other one was idle.) Tried everything, and here is what worked for me: set fan speed to Auto, check the fan curve to be idling at low temps (mine is below 40C) in fan settings, and then go to general settings and in the "Enable low-level hardware access interface" dropdown choose "kernel mode". My fans stopped revving immediately. Hopefully, it will fix my second issue too.

 

[Arjun Agarwal]

I am having the same exact issue, tried kernel mode but it did not help.. Someone save me!