How to quickly & easily fix coil-whine(coil choke noise)

[lexluthermiester]

Think I may see an issue. Maybe it was so thin it flattened/spread out and didn’t make enough contact with the board and choke? Maybe I have to build it up with another application?View attachment 286257

View attachment 286258
View attachment 286256

I don’t know. Some look like good contact though…

Yup, that's not enough. Enough glue needs to be used to completely saturate the spaces between the choke and the PCB.

How does too much super glue damage components if it is non conductive?

It doesn't. SuperGlue is both chemical inert and electrically insulative. It will NOT harm the electronics on the card.

This kind of glue does its "work" by dissolving a bit of the surfaces it enters in contact with.

No. That is absolute nonsense. Please read below;

Cyanoacrylate - Wikipedia

en.wikipedia.org

Note the section explaining the uses in electronics. If it were a solvent it would be completely inappropriate for use in electronics work.

 

[Wolverine2349]

Yup, that's not enough. Enough glue needs to be used to completely saturate the spaces between the choke and the PCB.


It doesn't. SuperGlue is both chemical inert and electrically insulative. It will NOT harm the electronics on the card.


No. That is absolute nonsense. Please read below;

Cyanoacrylate - Wikipedia

en.wikipedia.org

Note the section explaining the uses in electronics. If it were a solvent it would be completely inappropriate for use in electronics work.

So if SuperGlue does not damage components, how can you put too much on? You stated I had a bit much?? What is the harm then? Could it be less effective at eliminating coil whine? Though Stama says it would not and actually could help more with eliminating coil whine.

 

[lexluthermiester]

So if SuperGlue does not damage components, how can you put too much on? You stated I had a bit much??

Well yeah, but you don't want to encase your card in the stuff. With superglue, a little bit goes a long way, but sometimes you have to be generous. Just don't go over-board.

 

[stama]

It doesn't. SuperGlue is both chemical inert and electrically insulative. It will NOT harm the electronics on the card.

You are right, it's doesn't act by dissolving the materials. I was thinking that it does, because I knew it was not recommended to use it one certain materials (including the fiberglass PCBs are made of) due to releasing a lot of heat (exothermal reaction) but I see that it's not due to a chemical reaction, it's just what happens when these glues dry.

 

[Brothanumsie]

Edit: My mistake

 

[Veseleil]

still have no clue why closing that app creates more noise

Use FPS counter (fraps or alike), and find out.

 

[Wolverine2349]

…I honestly don’t believe it…it worked. Lex, you maniacal SOB!!!

i can hear it if I’m 2 inches away (this card is water cooled and in an open case, so no fan to cover the noise up), but normal playing position about 3 feet away, little to nothing. And that’s if the room is dead silent. Closing Heaven I can hear a little something (still have no clue why closing that app creates more noise). I’m going to do 1 more application and see if I can get it down to nothing but as is, I’m extremely happy. I can finally put some wattage through this monster.

I hope this helps anyone else thinking about giving this method a try. I can can only speak on the glue that I used which was extremely low viscosity. So multiple applications were needed but it’s a very clean finish.

So it worked finally. Did you let it dry longer or apply more super glue?

I am still letting mine dry. Been 36 hours. Though at midnight last night when it was drying for 24 to 25 hours, my power went out and temp in my house dropped from 67f to 60f. And power still not back on and in the 30s outside with inches of snow.

Is this going to affect the curing given lower house temp and lowrr humidity or was it fully cured at 24 to 25 hours just before my power went out?

 

[Brothanumsie]

So it worked finally. Did you let it dry longer or apply more super glue?

I am still letting mine dry. Been 36 hours. Though at midnight last night when it was drying for 24 to 25 hours, my power went out and temp in my house dropped from 67f to 60f. And power still not back on and in the 30s outside with inches of snow.

Is this going to affect the curing given lower house temp and lowrr humidity or was it fully cured at 24 to 25 hours just before my power went out?

So first application was 17 hours then tested. I applied more, let sit for 24 hours and tested this morning. It seems to dry quicker on itself. Much quicker than the first application

 

[Wolverine2349]

So first application was 17 hours the tested. I applied more, let sit for 24 hours and tested this morning. It seems to dry quicker on itself. Much quicker than the first application

Awesome so happy it worked for you. What card are you doing it on again?

 

[Brothanumsie]

Awesome so happy it worked for you. What card are you doing it on again?

6950XT liquid red devil

 

[stama]

@Brothanumsie, congratulations! If you have time, maybe you can do a new recording of the noise teh card makes in Heaven? I'd be curious to compare to how it was before.

@Wolverine2349 lower temps and lower humidity slow down the curing process, but bond strength improvement slows over time, the greatest improvement is at the beginning I think.

 

[SRV123]

…I honestly don’t believe it…it worked. Lex, you maniacal SOB!!!

i can hear it if I’m 2 inches away (this card is water cooled and in an open case, so no fan to cover the noise up), but normal playing position about 3 feet away, little to nothing. And that’s if the room is dead silent. Closing Heaven I can hear a little something (still have no clue why closing that app creates more noise). I’m going to do 1 more application and see if I can get it down to nothing but as is, I’m extremely happy. I can finally put some wattage through this monster.

I hope this helps anyone else thinking about giving this method a try. I can can only speak on the glue that I used which was extremely low viscosity. So multiple applications were needed but it’s a very clean finish.

Congrats!

Can you please specify which super glue you used? Also, when you re-applied did you completely seal up the gap between the PCB and the choke?

Also random shower thought but would probably not advise anyone to use a fan to blow on the PCB to speed up drying. I did it and now I'm wondering if that had interfered with the superglue settling properly on the PCB.

 

[Wolverine2349]

@Brothanumsie, congratulations! If you have time, maybe you can do a new recording of the noise teh card makes in Heaven? I'd be curious to compare to how it was before.

@Wolverine2349 lower temps and lower humidity slow down the curing process, but bond strength improvement slows over time, the greatest improvement is at the beginning I think.

Well my power came back 4 hours ago. The lowest temp it got was 61F in my house during that time with no heater and outside weather in the high 30s. Now temp back to normal 67F in my house.

Do you think it made any difference 6F lower temp? It had already been 25 hours of letting the card sit and dry before my power went out.

And do you think I need to wait any longer or go ahead and try it out?

 

[Brothanumsie]

UPDATE:
TLDR - It didnt work for me.

So ill do my best to explain. So this morning i put the card and waterblock back on but not the backplate, just to do a quick test. Thats when i assumed it worked since the whine was pretty much gone. I wasnt planning on using this system today so i took it back apart and applied a little more to hopefully eliminate any and all whine this time. I gave it about 8 hours and noticed it was completely dry, so i put everything back together again, backplate included. The whine was back. WTF!?!? The same noise, same intensity, same pitch. I was confused as hell. The only thing different was no backplate. I removed the backplate and gave it a shot...no whine. My hypothesis - The backplate standoffs and thickness relieves a small amount of pressure from block to chips on the front side. Without it, im guessing theres ever so slightly a tighter fit and the chokes are now touching the plating on that block, thus removing some vibration. So i screwed in all but 3 screws to secure the pcb to the waterblock and then screwed in the backplate on top of those screws to provide a little extra cooling on back. As you can see below, the chokes that i suspected were the issue, are now pretty snug against the block. (actually you probably cant see from this blurry ass picture, my bad)

So i cant say for sure if this method would work for me. Maybe some more drying time might help but everything looked pretty dry this morning and it made no difference. Maybe this is a more severe case and no amount of glue would help it but please dont let this discourage anyone else from trying. Sorry for the false alarm everyone.

 

[Wolverine2349]

UPDATE:
TLDR - It didnt work for me.

So ill do my best to explain. So this morning i put the card and waterblock back on but not the backplate, just to do a quick test. Thats when i assumed it worked since the whine was pretty much gone. I wasnt planning on using this system today so i took it back apart and applied a little more to hopefully eliminate any and all whine this time. I gave it about 8 hours and noticed it was completely dry, so i put everything back together again, backplate included. The whine was back. WTF!?!? The same noise, same intensity, same pitch. I was confused as hell. The only thing different was no backplate. I removed the backplate and gave it a shot...no whine. My hypothesis - The backplate standoffs and thickness relieves a small amount of pressure from block to chips on the front side. Without it, im guessing theres ever so slightly a tighter fit and the chokes are now touching the plating on that block, thus removing some vibration. So i screwed in all but 3 screws to secure the pcb to the waterblock and then screwed in the backplate on top of those screws to provide a little extra cooling on back. As you can see below, the chokes that i suspected were the issue, are now pretty snug against the block. (actually you probably cant see from this blurry ass picture, my bad)

View attachment 286428

So i cant say for sure if this method would work for me. Maybe some more drying time might help but everything looked pretty dry this morning and it made no difference. Maybe this is a more severe case and no amount of glue would help it but please dont let this discourage anyone else from trying. Sorry for the false alarm everyone.

So wait was it the backplate applied a certain way that eliminated the coil whine??

 

[Brothanumsie]

So wait was it the backplate applied a certain way that eliminated the coil whine??

Nope. It’s NOT applying the backplate that allowed a tighter fit. Which I suspect eliminates the vibration.

 

[Wolverine2349]

Nope. It’s NOT applying the backplate that allowed a tighter fit. Which I suspect eliminates the vibration.

Oh so not applying the backplate? Did you ever try no backplane before applying the super glue?

Is it possible to run card without the backplate and thus maybe no whine?

 

[SRV123]

Nope. It’s NOT applying the backplate that allowed a tighter fit. Which I suspect eliminates the vibration.

So you're saying you:

1. Disassembled the gpu completely until bare PCB (with backplate removed)
2. Applied superglue on inductors
3. Waited till everything dried
4. Reassembled GPU completely (backplate included)

Because if so then it might be better to assemble the GPU back together completely immediately after having applied the super glue.

 

[Wolverine2349]

So you're saying you:

1. Disassembled the gpu completely until bare PCB (with backplate removed)
2. Applied superglue on inductors
3. Waited till everything dried
4. Reassembled GPU completely (backplate included)

Because if so then it might be better to assemble the GPU back together completely immediately after having applied the super glue.

Yeah interesting point though I am not sure do the thermal pads make contact with the chokes?

Do they make contact with chokes on the PNY 4090?

 

[Brothanumsie]

Yeah interesting point though I am not sure do the thermal pads make contact with the chokes?

Do they make contact with chokes on the PNY 4090?

No thermal pads on chokes

 

[Wolverine2349]

Sad to say no difference in the PNY RTX 4090 after the Super glue method. Still same whine/buzz noise.

Even tried removing the backplate to see if it helped and nope no difference at all.

There is a small R2 inductor on backplate side of PCB and I tried putting my finger on it to see if noise was coming from it and it improved and nope. Even tried covering up the MLCC behind the GPU and no difference.

Buzz/whine most coming form front of PCB I think.

This was after I let it dry for almost 48 hours.

Card still works and will be sold.

Gaming OC is next stop. Seems PCB it is easier to apply super glue and wick it as inductors spaced farther apart.

 

[athenaesword]

UPDATE:
TLDR - It didnt work for me.

So ill do my best to explain. So this morning i put the card and waterblock back on but not the backplate, just to do a quick test. Thats when i assumed it worked since the whine was pretty much gone. I wasnt planning on using this system today so i took it back apart and applied a little more to hopefully eliminate any and all whine this time. I gave it about 8 hours and noticed it was completely dry, so i put everything back together again, backplate included. The whine was back. WTF!?!? The same noise, same intensity, same pitch. I was confused as hell. The only thing different was no backplate. I removed the backplate and gave it a shot...no whine. My hypothesis - The backplate standoffs and thickness relieves a small amount of pressure from block to chips on the front side. Without it, im guessing theres ever so slightly a tighter fit and the chokes are now touching the plating on that block, thus removing some vibration. So i screwed in all but 3 screws to secure the pcb to the waterblock and then screwed in the backplate on top of those screws to provide a little extra cooling on back. As you can see below, the chokes that i suspected were the issue, are now pretty snug against the block. (actually you probably cant see from this blurry ass picture, my bad)



So i cant say for sure if this method would work for me. Maybe some more drying time might help but everything looked pretty dry this morning and it made no difference. Maybe this is a more severe case and no amount of glue would help it but please dont let this discourage anyone else from trying. Sorry for the false alarm everyone.

hi i read your post a couple of times. Did you manage to eliminate the whine after messing around with the backplate mounting pressure via screws adjustment?
specifically :"So i screwed in all but 3 screws to secure the pcb to the waterblock and then screwed in the backplate on top of those screws to provide a little extra cooling on back. "
what did that do for noise?

are you using an ekwb block with active back plate or just the black one? i don't think that card can run without the backplate on so just clarifying your set up since you managed to test the card without the BP on.

 

[Wolverine2349]

hi i read your post a couple of times. Did you manage to eliminate the whine after messing around with the backplate mounting pressure via screws adjustment?
specifically :"So i screwed in all but 3 screws to secure the pcb to the waterblock and then screwed in the backplate on top of those screws to provide a little extra cooling on back. "
what did that do for noise?

are you using an ekwb block with active back plate or just the black one? i don't think that card can run without the backplate on so just clarifying your set up since you managed to test the card without the BP on.

Also did they put more super glue on before putting backplate back on as their post seemed to indicate? Could that have caused problems where original amount of superglue was good?

 

[Brothanumsie]

hi i read your post a couple of times. Did you manage to eliminate the whine after messing around with the backplate mounting pressure via screws adjustment?
specifically :"So i screwed in all but 3 screws to secure the pcb to the waterblock and then screwed in the backplate on top of those screws to provide a little extra cooling on back. "
what did that do for noise?

are you using an ekwb block with active back plate or just the black one? i don't think that card can run without the backplate on so just clarifying your set up since you managed to test the card without the BP on.

No active backplate, just the red devil backplate. I didnt try to adjust the tightness of the screws, just the normal amount we usually screw them in (not too tight).

Also did they put more super glue on before putting backplate back on as their post seemed to indicate? Could that have caused problems where original amount of superglue was good?

So i dont think the glue had anything to do with this entire situation. I did add more glue after my test this morning where the card was silent without the backplate. After putting more glue on, waiting for it to dry, then putting everything back together including the backplate, the whine was back same as ever. Im almost certain its got to do with the slight pressure adjustment from having a backplate, to just pcb and waterblock. The chokes have a little more pressure on them now and thats stopping some vibration. Im no engineer so i cant prove my theory but process of elimination tells me the glue didnt do much.

Edit: So i noticed there are rubber gromets below the standoffs of the backplate. My guess is with the backplate, the screws stop at the top of the standoffs. Without it, im able to apply just slightly more pressure, pushing down on those gromets. Again, just a theory

 

[lexluthermiester]

I am still letting mine dry. Been 36 hours.

So first application was 17 hours then tested. I applied more, let sit for 24 hours and tested this morning.

There is no reason to wait that long. The application is usable within 30 minutes to an hour, tops.

It seems to dry quicker on itself. Much quicker than the first application

True, this does happen.

UPDATE:
TLDR - It didnt work for me.

So ill do my best to explain. So this morning i put the card and waterblock back on but not the backplate, just to do a quick test. Thats when i assumed it worked since the whine was pretty much gone. I wasnt planning on using this system today so i took it back apart and applied a little more to hopefully eliminate any and all whine this time. I gave it about 8 hours and noticed it was completely dry, so i put everything back together again, backplate included. The whine was back. WTF!?!? The same noise, same intensity, same pitch. I was confused as hell. The only thing different was no backplate. I removed the backplate and gave it a shot...no whine. My hypothesis - The backplate standoffs and thickness relieves a small amount of pressure from block to chips on the front side. Without it, im guessing theres ever so slightly a tighter fit and the chokes are now touching the plating on that block, thus removing some vibration. So i screwed in all but 3 screws to secure the pcb to the waterblock and then screwed in the backplate on top of those screws to provide a little extra cooling on back. As you can see below, the chokes that i suspected were the issue, are now pretty snug against the block. (actually you probably cant see from this blurry ass picture, my bad)

View attachment 286428

So i cant say for sure if this method would work for me. Maybe some more drying time might help but everything looked pretty dry this morning and it made no difference. Maybe this is a more severe case and no amount of glue would help it but please dont let this discourage anyone else from trying. Sorry for the false alarm everyone.

That's just weird. The backplate shouldn't have anything to do with it...