Hardware GPU Fan makes weird clicking noise

[ivt18]

So, since about a week ago I have noticed my GPU fan makes a weird rattling noise whenever it's spinning. It doesn't seem to depend on the RPMs, and it just comes and goes every now and then, but it does seem to have gotten louder over the last few days. This morning I checked after looking at a couple of forums wheter the fan was loose and if it could spin freely, and it all seems fine to me. By the way, my GPU is a MSI ARMOR RX 570 4GB OC. I don't know if these cards are notorious for this, but I'm pointing it out just in case. I also have a kinda aggressive fan curve. I would be very glad if I could get some help.
I can't seem to be able upload an audio file from my phone, but I found a YouTube video where the GPU makes pretty much the exact same noise as mine.

 

[raystriker]

It's possible that the fan is just slightly off axis?

 

[FAST6191]

Usual suspects are
1) dust or something (say a wire from something else) in the path of the blades. Blow it out/move it out.
2) The blade section stuffed on the motor have come loose (either going at an angle or rising up/lowering down) and is hitting against something. See if you can sort it -- most computer fans are just pressed onto a spinning shaft rather than keyed or pinned to a flat or something which makes life easier, though some will also bend things out the way or file down a bit of the fan blade assembly (minor differences in shape can make major differences in efficiency, noise or whatever, and we can have a nice month long lecture series on fluid dynamics for fans if you want, but at the same time meh for a fan like this if you are only taking something off with a needle file).
3) A bearing somewhere in the mix is broken. Can mean replacement (which probably means replacement fan assembly, or rigging up some kind of replacement setup from something you might be more easily able to buy) or tearing it down far enough (might be a simple seal, might be something more*) that you can get a drop of oil in there -- try to know what you are doing there as squirting oil around the place is usually less than ideal.

1) is most common and easiest to fix. The others are failures at some level and can be harder to sort, you can however usually tell what goes and possible indicate to yourself if you apply a bit of pressure during operation (pokey stick, screwdriver, finger... I have used all of them). After things "get to temperature" as it were in the assembly then it might stop, or might even get worse.

Resonance I would usually expect to be there from the start but a motor on the way out and some dust might change the harmonics such that it hits it. Not to mention you say this happens at any speed.

*this is for a desk fan but I have seen similar setups inside computers on a smaller scale.

 

[ivt18]

Usual suspects are
1) dust or something (say a wire from something else) in the path of the blades. Blow it out/move it out.
2) The blade section stuffed on the motor have come loose (either going at an angle or rising up/lowering down) and is hitting against something. See if you can sort it -- most computer fans are just pressed onto a spinning shaft rather than keyed or pinned to a flat or something which makes life easier, though some will also bend things out the way or file down a bit of the fan blade assembly (minor differences in shape can make major differences in efficiency, noise or whatever, and we can have a nice month long lecture series on fluid dynamics for fans if you want, but at the same time meh for a fan like this if you are only taking something off with a needle file).
3) A bearing somewhere in the mix is broken. Can mean replacement (which probably means replacement fan assembly, or rigging up some kind of replacement setup from something you might be more easily able to buy) or tearing it down far enough (might be a simple seal, might be something more*) that you can get a drop of oil in there -- try to know what you are doing there as squirting oil around the place is usually less than ideal.

1) is most common and easiest to fix. The others are failures at some level and can be harder to sort, you can however usually tell what goes and possible indicate to yourself if you apply a bit of pressure during operation (pokey stick, screwdriver, finger... I have used all of them). After things "get to temperature" as it were in the assembly then it might stop, or might even get worse.

Resonance I would usually expect to be there from the start but a motor on the way out and some dust might change the harmonics such that it hits it. Not to mention you say this happens at any speed.

*this is for a desk fan but I have seen similar setups inside computers on a smaller scale.

I cleaned my entire PC a couple weeks ago with a can of compressed air, so I doubt it's number 1. There are also no visible wires. Regarding number 2 though, maybe it has bent down? My GPU is mounted vertically because the case only allows for it to be mounted that way (Thermaltake Core G3), so maybe that's the case. I really hope it's not number 3, even though I'm pretty sure I still have my warranty, but I'd still prefer if I could sort it out myself.
I also tried this morning to push down the fan to make sure it wasn't loose. I'll keep troubleshooting tonight.

--------------------- MERGED ---------------------------

It's possible that the fan is just slightly off axis?

That's what I initially thought, but at a glance it looks as straight as the other fan which doesn't make the noise.

 

[ivt18]

So this morning I took another look at it and I found that the left fan is just fine, but the right fan behaves almost like a button when I press it, like if it was spring loaded. I can't seem to upload a video of it so you will just have to take my word for it.
Does this behaviour mean that the bearing is going?

 

[ivt18]

In the end, I conclude that when I was cleaning my PC a couple weeks ago with compressed air I overspun the GPU fan, thus damaging the bearing. I just read this on a Reddit post and honestly, it makes sense. Lesson learnt: hold fans in place when cleaning.

 

[FAST6191]

That would be an impressive amount of spin, more than I have seen from the average canned air.

Spinning is mostly considered a problem because few motors in computers seem to include a flyback diode so when they are spun they act as generators and pump voltage back in -- this is why the suggestions for cleaning are either stick a toothpick in there to prevent it spinning OR unplug it from the computer.

 

[ivt18]

That would be an impressive amount of spin, more than I have seen from the average canned air.

Spinning is mostly considered a problem because few motors in computers seem to include a flyback diode so when they are spun they act as generators and pump voltage back in -- this is why the suggestions for cleaning are either stick a toothpick in there to prevent it spinning OR unplug it from the computer.

I don't think in this case it has anything to do with the motor pumping out voltage since the fan itself seems to have a hardware problem. I'm uncertain I can return the GPU and get it fixed, since it's only 1 year old, but I'm guessing they would count it as me damaging the part. I guess I will deal with it for now and in the future I might upgrade to a NZXT Kraken G12 with an AIO, or just upgrade the GPU altogether. Thanks for the help anyways!

 

[The Real Jdbye]

That would be an impressive amount of spin, more than I have seen from the average canned air.

Spinning is mostly considered a problem because few motors in computers seem to include a flyback diode so when they are spun they act as generators and pump voltage back in -- this is why the suggestions for cleaning are either stick a toothpick in there to prevent it spinning OR unplug it from the computer.

I have heard many say that it can damage the fan if you don't hold the blades in place when you clean them. I was never sure how true that was, but I guess they were right.
Don't motors need to have a diode in line anyway, to prevent a power surge feeding back to the electronics when the motor stops? I thought that was pretty much a requirement.

In the end, I conclude that when I was cleaning my PC a couple weeks ago with compressed air I overspun the GPU fan, thus damaging the bearing. I just read this on a Reddit post and honestly, it makes sense. Lesson learnt: hold fans in place when cleaning.

I've vacuumed fans and let them spin up to super speed a few times just for fun. I stopped doing that once I heard it could damage the fan, though. At least now I know I wasn't being careful for nothing.
At least it's an older cheap GPU, it's probably about due for a replacement anyway.

 

[ivt18]

I have heard many say that it can damage the fan if you don't hold the blades in place when you clean them. I was never sure how true that was, but I guess they were right.
Don't motors need to have a diode in line anyway, to prevent a power surge feeding back to the electronics when the motor stops? I thought that was pretty much a requirement.

I've vacuumed fans and let them spin up to super speed a few times just for fun. I stopped doing that once I heard it could damage the fan, though. At least now I know I wasn't being careful for nothing.
At least it's an older cheap GPU, it's probably about due for a replacement anyway.

Yeah, still kind of a bummer though, as the card performs great. For now the temperatures seem to be fine, but if they start rising because of the fan not working properly, I'll either bite the bullet and let the card thermal throttle, or upgrade to a Sapphire Nitro+ 5500XT, as it seems to be a pretty good GPU for the money.
EDIT: I have actually found the RX 590 performs better, however, I cannot seem to find one from a reliable source at a good price and that ships to Spain. F