Strix RTX 3090 Fan vibration issues

[pk1209]

I'm honestly shocked that people are putting something as flammable as paper towel directly onto something that is designed to get to 80 degrees and beyond! Surely a bit of silicone or even some folded up electrical tape would make more sense and have the same affect?

[KMagic]

I just got a Strix 3090 Christmas week (purchased from newegg with next day air shipping). I had already been researching the card and had found this thread so I was eager to get my card in and test it. Its now been just shy of 2 weeks in my system, and I have tested the card extensively, stress tested it, overclocked it, and put some heavy use on it. I have had NO fan noise issue, no vibration as mentioned in this thread, or like I have seen in several videos. The card does have some fairly strong coil whine however, that could almost mimic the sound of a vibrating fan to someone who isnt familiar with these issues, but is 100% coil whine noise.

I wanted to post however because I noticed something with my card that MAY be a fix for this issue already being implemented - but I need some response here to know for sure. I noticed when I took the card out of the packaging that my fan shroud is very loose. Like I can grip it on either side and physically slide it around. It is not super stiff like every other video card I have owned in the past. The movement is only in the shroud itself - the fans themselves do not move with the shroud, and the heatsink is rock solid, its just the plastic shell, and it appears very deliberately designed this way. The range of movement isnt a ton, but its enough that I immediately noticed it when I picked it up. It appears from all I can tell however that it was designed this way. It has enough play in it however that with my vga support bracket on the end of the card, it pushes the shroud up on that end, and almost makes the card look slightly crooked in the socket, as if Im pushing up too hard on the card, but its just because of the slight play in the shroud makes it appear that way.. again Im not talking about massive range of movement here but at least a couple millimeters of play. I want to know if anyone can verify how solid your shroud feels on your strix 3090? Is it super sturdy and rigid or does it have a little play in it?

I have tried pressing and moving the shroud as the card is running full tilt and I cant make it produce any fan vibration noise. My card seems to not have this issue at all. Hopefully it doesn't develop it, but I've put hours and hours of stress testing on it so far, with fans running between 70-95% the whole time. I have of course forced other fan speeds under load to look for the issue, and its simply not present in my 2 week old card.

EDIT: also just to add for those who may be curious - my card had plastics to remove EVERYWHERE. Lightbar, front face plate, backplate, and may have even been on the fan motors but I dont recall 100% if it was on the fans. I add this because many mentioned not having plastics to peel off.

[Megascrypt]

I just got a Strix 3090 Christmas week (purchased from newegg with next day air shipping). I had already been researching the card and had found this thread so I was eager to get my card in and test it. Its now been just shy of 2 weeks in my system, and I have tested the card extensively, stress tested it, overclocked it, and put some heavy use on it. I have had NO fan noise issue, no vibration as mentioned in this thread, or like I have seen in several videos. The card does have some fairly strong coil whine however, that could almost mimic the sound of a vibrating fan to someone who isnt familiar with these issues, but is 100% coil whine noise.

I wanted to post however because I noticed something with my card that MAY be a fix for this issue already being implemented - but I need some response here to know for sure. I noticed when I took the card out of the packaging that my fan shroud is very loose. Like I can grip it on either side and physically slide it around. It is not super stiff like every other video card I have owned in the past. The movement is only in the shroud itself - the fans themselves do not move with the shroud, and the heatsink is rock solid, its just the plastic shell, and it appears very deliberately designed this way. The range of movement isnt a ton, but its enough that I immediately noticed it when I picked it up. It appears from all I can tell however that it was designed this way. It has enough play in it however that with my vga support bracket on the end of the card, it pushes the shroud up on that end, and almost makes the card look slightly crooked in the socket, as if Im pushing up too hard on the card, but its just because of the slight play in the shroud makes it appear that way.. again Im not talking about massive range of movement here but at least a couple millimeters of play. I want to know if anyone can verify how solid your shroud feels on your strix 3090? Is it super sturdy and rigid or does it have a little play in it?

I have tried pressing and moving the shroud as the card is running full tilt and I cant make it produce any fan vibration noise. My card seems to not have this issue at all. Hopefully it doesn't develop it, but I've put hours and hours of stress testing on it so far, with fans running between 70-95% the whole time. I have of course forced other fan speeds under load to look for the issue, and its simply not present in my 2 week old card.

EDIT: also just to add for those who may be curious - my card had plastics to remove EVERYWHERE. Lightbar, front face plate, backplate, and may have even been on the fan motors but I dont recall 100% if it was on the fans. I add this because many mentioned not having plastics to peel off.

I can confirm your findings. At first I didn't think anything of it, but now since you've mentioned, there is a slight shroud movement as if it's not tight enough.

I also got my card at Newegg and the whole card incl. the centre of the fans had plastic protective film. Unfortunately my card has fairly loud vibration sound regardless of vertical or horizontal placement.

I have 35 days left to return the card to Newegg for full refund or replacement. I don't want to end up without card so for now I'm still keeping it, but if I won't be able to get another strix 3090 from Newegg within 35 days as the stock isn't there I will return my current Strix and probably move to different AIB I don't want to end up with faulty card design since ASUS seem to ignored that and we, effected customers have been put on the back burner

BTw. This is my first Asus card and from the looks of it, it might be my last.

[garfieldsorlando]

I can confirm your findings. At first I didn't think anything of it, but now since you've mentioned, there is a slight shroud movement as if it's not tight enough.

I also got my card at Newegg and the whole card incl. the centre of the fans had plastic protective film. Unfortunately my card has fairly loud vibration sound regardless of vertical or horizontal placement.

I have 35 days left to return the card to Newegg for full refund or replacement. I don't want to end up without card so for now I'm still keeping it, but if I won't be able to get another strix 3090 from Newegg within 35 days as the stock isn't there I will return my current Strix and probably move to different AIB I don't want to end up with faulty card design since ASUS seem to ignored that and we, effected customers have been put on the back burner

BTw. This is my first Asus card and from the looks of it, it might be my last.

It is definitely a design flaw. After putting some shim to lift fans a bit away from the heatsink, vibration noise is gone. I am holding a bit with RMA with newegg as the coil whine for my unit is minimal. Trying to get another one before I sent this one back.

[curthard89]

I just got a Strix 3090 Christmas week (purchased from newegg with next day air shipping). I had already been researching the card and had found this thread so I was eager to get my card in and test it. Its now been just shy of 2 weeks in my system, and I have tested the card extensively, stress tested it, overclocked it, and put some heavy use on it. I have had NO fan noise issue, no vibration as mentioned in this thread, or like I have seen in several videos. The card does have some fairly strong coil whine however, that could almost mimic the sound of a vibrating fan to someone who isnt familiar with these issues, but is 100% coil whine noise.

I wanted to post however because I noticed something with my card that MAY be a fix for this issue already being implemented - but I need some response here to know for sure. I noticed when I took the card out of the packaging that my fan shroud is very loose. Like I can grip it on either side and physically slide it around. It is not super stiff like every other video card I have owned in the past. The movement is only in the shroud itself - the fans themselves do not move with the shroud, and the heatsink is rock solid, its just the plastic shell, and it appears very deliberately designed this way. The range of movement isnt a ton, but its enough that I immediately noticed it when I picked it up. It appears from all I can tell however that it was designed this way. It has enough play in it however that with my vga support bracket on the end of the card, it pushes the shroud up on that end, and almost makes the card look slightly crooked in the socket, as if Im pushing up too hard on the card, but its just because of the slight play in the shroud makes it appear that way.. again Im not talking about massive range of movement here but at least a couple millimeters of play. I want to know if anyone can verify how solid your shroud feels on your strix 3090? Is it super sturdy and rigid or does it have a little play in it?

I have tried pressing and moving the shroud as the card is running full tilt and I cant make it produce any fan vibration noise. My card seems to not have this issue at all. Hopefully it doesn't develop it, but I've put hours and hours of stress testing on it so far, with fans running between 70-95% the whole time. I have of course forced other fan speeds under load to look for the issue, and its simply not present in my 2 week old card.

EDIT: also just to add for those who may be curious - my card had plastics to remove EVERYWHERE. Lightbar, front face plate, backplate, and may have even been on the fan motors but I dont recall 100% if it was on the fans. I add this because many mentioned not having plastics to peel off.

The shroud very much moves more than most, but seems to be a free floating thing (to an extent), mine does not move side to side, but can compress on both ends (how i've managed to stop mostly stop the noise - with a sag holder both ends pushing both ends up)

Happens at 80-100% fan, but also massively temp dependant. So just waiting for Asus to actually help us out instead of concentrating on getting new cards out and not caring about us. Im a patient man, its not impacting my enjoyment of gaming (headphones etc), but time will tell, just want it sorted :-)

What kind of temps are you getting? Mine will happily reach 70deg with 72% fan with 100% load and 100% power draw, some games do 100% load but only like 80% power draw so sits in the low 60's. Most unreal engine games for unknown reasons will draw absolutely everything, stick the card at 123%, they will draw 123% and just roast the card (obviously clocks drop when power limited), i normally leave mine around 110% (i undervolt to reduce power and sustain clocks more)*

[gooper23]

I'm honestly shocked that people are putting something as flammable as paper towel directly onto something that is designed to get to 80 degrees and beyond! Surely a bit of silicone or even some folded up electrical tape would make more sense and have the same affect?

1. The copper heat pipes are nowhere near 80*C
2. 80*C does not cause paper to combust.
3. I agree there are alternative materials. I used a cut up piece of double sided tape that seemed to fit just right.

[KMagic]

Well based on the replies it seems my theory has been shot down. I was thinking the loose shroud was maybe something new that asus had done to address this issue, but it appears owners of older cards also have the same loose shroud. Based on the videos I had seen it appeared to me that the shroud was very firm and stiff. If they are all loose like mine however, then nothing has changed. I guess all I can add to this discussion then is to say that this issue does not affect every card, because I do not have this problem with vibration noises related to the fans, and I've been paying very close attention to the card the past 2 weeks as I have been stress testing and overclocking.

Regarding my temps - I run the card currently at a power limit of 107%.. In my testing of this card the sweet spot is between 105% and 110% power limit. This puts the card drawing around 400-420w of power. In my testing, taking the power limit beyond 110%, in a lot of games, results in a miniscule additional boost of performance. I spent hours and hours and hours testing the power limit, benchmarking, changing the limit, benchmarking, changing it again and benchmarking, and noting temperature and power draw at each step. Running the card at the 123% max power limit at this point is really really pointless in my opinion. You increase the temperature of the card by several degrees immediately, and often gain nothing tangible for it. This is because, as I mentioned, the card will often boost to 2000mhz+ at a 107% power limit. My card starts to show instability when you get over 2055mhz. It may run one game fine hitting 2055mhz or even 2070mhz occasionally, but then crash 20 minutes into a different game, so trying to push above 2040mhz. The way the GPU boost works makes it VERY hard to say "I want to shoot for 2055mhz" Because you can get one game to boost to 2055mhz and be ok, then lanch another game, with a different load, and the gpu will try to boost to 2070mhz, which may cause a crash. If you adjust the core down to keep it from going over 2055mhz in that game, then it drops to 2025mhz in the other game. I spent several days perfecting a custom voltage/clock curve, only to eventually, after LOTS of testing, throwing the thing out and just going with a flat +60 on the core. +60 on the core will not crash in anything Ive tested. Some games I will hit 1980mhz, other games Ill see 2040mhz, its all in the different type of load and how the card responds. If you push too hard on the core, then you find that some games will try to boost way too high and cause you to crash. The only way to control this is to find the highest offset you can apply that will work with everything. +75mhz in my testing is about the limit for my strixx 3090 OC edition. The next step up would be +90mhz, and this will work most everywhere but the closer I got to a 100mhz offset the more instability I would eventually start hitting.

So, the card will already clock over 2000mhz at 110% power limit, so pushing the power limit to 123% will only cause the average clock over a given time period to be higher. For example, if you ever run 3d mark and watch your clocks in real time, they bounce around a lot along the voltage/clock curve. Pushing the power limit up to the max will enable the card to spend more time in the upper ends of that curve, so instead of dropping from say 2040mhz down to 1995mhz in one section of a benchmark, it may only drop down to 2010mhz instead, so at the end of the benchmark your average clock is a little higher. It nets a few more points in 3d mark scores, and costs about 7 or 8 degrees higher temperature, which is not at all worth it. GPU boost is already so aggressive that overclocking video cards these days is a lot more tricky than it used to be back when clocks were just a fixed value. I have even had moments where a certain overclock is unstable, and you can run the same benchmark 20 times in a row and it will crash on the 21st time as soon as the card tries to hit 2070mhz or something, eventhough it bossted to it fine 20 times prior. It means that 2070mhz isnt 100% stable, but given the card only sits there for a few seconds typically before dropping back down, you can run some quick tests and think you are stable only to crash hours later. You have to really be cautious on the overclocks for this reason, and is why I finally found +60 and 105 - 110% power limitto be a super solid sweet spot. The card will boost to its near maximum stable clock of ~2040mhz in some games, but on average sits around 2000mhz, and keeps the temperatures in my case right at 70c max load with fans around 75 to 80%. If I max fans out to 100% I can keep the card around 68/69c under full long duration load. The case cooling set up has a big impact on this though. I am moving to a new Lian Li o11 dynamic XL build this weekend, and out of a Phanteks p400a build, so I expect to see my gpu temps come down. I will have intakes on the floor of the case feeding into the GPU, and wont have a CPU radiator blowing warm air into the card.

I rambled a lot here but maybe someone finds it useful. This has kinda been consuming my free time the last 2 weeks doing all this testing so.

[pk1209]

Blake just told me to RMA my Card (EU).... i dont think i will do that, theres almost no stock and i cant wait like 4 Weeks for my card

Yeah, same for me.. but RMA process doesn't even recognise my serial number, so can't even proceed if I wanted to. Have emailed and asked about advanced replacement as an option, as not going to sit without a card for weeks on end... Shouldn't have to amount of money I spent on this thing

[pk1209]

Reply from my email to Asus:

"Thank you for contacting ASUS Support. My name is Arsen

Please be advised that the RMA/return process is via the place of purchase only, true to the distribution channel. There is no direct RMA process with ASUS, this is why the form is not accepting the serial number.

I understand this is inconvenient, however in this case we suggest you contact your retailer for advice on returns/replacement or any other options they can offer. As part of the sale process the retailer accepted that they would control all returns/replacement on behalf of ASUS for the duration of the warranty. They should be able to sort this out for you. ASUS has no other policy on these devices; We cannot replace or repair these devices.

If you have any further questions feel free to contact us again."

Blade said I could do it through Asus, Asus say I can't... @Blake - can you advise here. No point RMA'ing with retailer if there is no stock

[curthard89]

Reply from my email to Asus:

"Thank you for contacting ASUS Support. My name is Arsen

Please be advised that the RMA/return process is via the place of purchase only, true to the distribution channel. There is no direct RMA process with ASUS, this is why the form is not accepting the serial number.

I understand this is inconvenient, however in this case we suggest you contact your retailer for advice on returns/replacement or any other options they can offer. As part of the sale process the retailer accepted that they would control all returns/replacement on behalf of ASUS for the duration of the warranty. They should be able to sort this out for you. ASUS has no other policy on these devices; We cannot replace or repair these devices.

If you have any further questions feel free to contact us again."

Blade said I could do it through Asus, Asus say I can't... @Blake - can you advise here. No point RMA'ing with retailer if there is no stock

Also got message from Blake, rather confusing.

No point of sending back to retailer until they have excess stock (which won't happen for a long while yet), not really sure what to do, also, if they won't repair and get new cards, I want my specific card back, I know its exact traits and clocks and voltages it runs at, don't want something with is subpar to what I have now (which is possible).

I've asked the retailer I bought my card from what the next steps are — I shall report back with information when I have it.