https://turingpi.com logo
Join Discord
Powered by
Turing Pi + PicoPs + cm4 = high pitch noise?
# │forum
g
It seems to be the case that my raspberry pi compute modules make a high pitch noise when powered on. I’m starting to suspect this may have something to do with using the pico PSU. Is anybody else experiencing this? Has anybody been any to test their Turing board + compute module with a different power supply and see if it still makes noise? I don’t have things necessary to do this unfortunately. Curious about others experiences.
d
I have a PicoPSU and CMs running next to me. I hear absolutely nothing even if I put my ear next to them.
g
For me it’s loud enough you can hear it without even being within 10 feet of the unit. Will be curious to see if anybody else has this.
d
Can you locate the sound?
g
I haven’t yet because I’ve been busy but I plan to
But the sound does not happen if the Pi’s are powered off
m
probably a whine coil, possible sources: - power supply - GPU - SSD
d
Yeah, this sounds like a coil whine. There are coild both on the PicoPSu and the TPi2 board near the nodes.
g
No SSD in it and no GPU based devices
It’s just a TPI2 and 2 compute modules
u
its not just the Pico, mine does this with a full ATX power supply
g
Oh that’s interesting
Raspberry pi compute modules?
u
I suspect its the TPI itself, probably the coils on the power supply for each node...
to be clear though, its not dangerous at all... its just power doing power things 🙂
d
And coils doing coil things
And this sound means some work is being done 🙂
g
It’s not so much that it’s dangerous as it’s super annoying
c
If is of any consolation I have the same and is launder when there is a bit of load. If the CM4 is completely idle I don’t hear it
a
All of my 3 CM4s whine very loud when powered on. But it’s definitely them (or the adapter boards) very annoying for me too. It’s louder then having fans next to me
m
you could try to hot glue the components making the noise to reduce vibration but that's about it
e
hey all, same here today delivery, turing2 + 1x CM4 and 3xSOQUARTZ pitch noice starts almost from start on single mode and on multiple as well. And yes in super idlle there is no noice ot its a bit lower. power via picoPSU and 'proper' PSU from turing shop.
u
Add me to the list. Sounds like coil whine but there's no change in pitch with operation.
Not coming from the Pico, seems to be coming directly from the tpi2.
Only occurs while powering CM4 as others have reported.
d
It won't change pitch since there's no driver to control the power delivery
u
Might not be an issue if placed in an enclosed ATX case but very audible on bench stand.
g
I have mine in a case and you can still hear it easily
c
I have the same effect and if i plug the cm4 in an other board there is no noise so, its the board or the board in combination with the power supply 😦
g
Thanks. I was hoping somebody with another board to plug it into could test
c
Check also an other power supply, so its really the board. 😦
c
Another board meaning something that is not a Turing pi 2?
c
Yes an non Turing pi 2 carrier board.
g
It seems like some people don’t notice it or it’s not happening. It may not be all boards but it sure seems like at least some have this issue
For mine it’s loud enough there’s no way you wouldn’t notice it
m
I get a pitch noise too, it seemed that each of the 4 slots produce this noise, I am not sure if it is the adapter boards or some component on the tpi itself near the slot.
g
I wonder what the cause is and if there is any way to stop it
Or if those of us who have it are just stuck with this loud horrible noise
b
I have the same problem. It is really annoying, even in a case. Louder than the fan.
g
I’m hoping somebody eventually figures out how to quiet it
Or at least narrows down if it’s the socket, adapter board, or pi itself
I suspect it’s the adapter board
f
I have the same issue, the same CM4s don't do this in a PiKVM PCIe or a Waveshare Raspberry Pi I/O board (but this has a horribly loud fan). I don't hear it when the case is closed though
I figured this when I got the device a few weeks ago, and used a toilet roll to locate it btw 😄
g
Thanks fnord. It seems to be a pretty common problem
f
yeah but can you workaround it? for example by using some fans in a case and closing the case?
g
For me personally, I have it in a case and you can still hear it across the room. It seems to have different strength for different people
Or perhaps maybe some are just more sensitive to the sound than others I don’t know
f
younger people and animals can be very sensitive yeah
g
I’m def not younger people
m
if you can find the exact source, you could hot glue the part that is vibrating
usually it's a toroidal power transformer
g
wouldn't that be like a coiled electromagnet?
i havent' seen one of those on the CMP4 board or its adapter
which seems to be, from what we have all had happen, the source of the issue
j
I have the same exact problem 😭 been using RPi4s with PoE hats previously and they could get noisy but at least you can adjust the speeds on those. This is even more annoying, makes me want to hide the TPi away…
g
Yeah. For me it’s loud enough that it basically makes me not want to use the tpi2 unless it’s very very far away
u

https://cdn.discordapp.com/attachments/1097526596040138893/1131208045192560681/IMG_20230713_162340452.jpg▾

https://cdn.discordapp.com/attachments/1097526596040138893/1131208045884616795/IMG_20230713_162343560.jpg▾

https://cdn.discordapp.com/attachments/1097526596040138893/1131208046304043118/IMG_20230617_202440748.jpg▾

applying dowsil 744 silicone sealant substantially cut down the amount of coilwhine
More accurate application might be cleaner and more successful than what I managed, but I got it down to perfectly tolerable levels even though its going to live in my living room
I got a 90ml tube of this
it was more than enough
I did find it slightly unwieldy so make sure to cover things it shouldn't touch, like the compute module slots, the fan headers, etc.
s
Same for me - starts when powering a node on, stops when all are off. Nodes: SOQUARTZ + RADXA CM3 on CM4 adapter board(s)
g
Which parts of the board specifically should I be trying to cover?
u
The ones that make noise which seems to correspond to the inductor coils. If it's only the adapter boards then the three orange squares might be all you need
g
Ok got it. I’ll try to buy some and do the 3 orange squares and see what happens
Thank you 🙏🏻
u
it takes about a day to harden, but I applied it over a few rounds with days (or weeks lol) in-between because I noticed a few other components still making noise
its a pretty stiff compound even when its not hardening, but I mostly managed to get results just squeezing it out of the tube
m
which component is the inductor coil? i tried putting red insulating varnish on grey boxes but i still hear noise on a slight cycle
r
I'm hearing this big time with the off-brand picoPSU shipped with the board. I'm going to try an official one and see if that helps.
Tried powering up all 4 CM4's with a proper Mini-box PicoPSU-150-XT. There is still noticeable whine but it sounds different. Not a fix :/
@Usandru I guess I'm going to have to use your silicone method on the carrier boards. How has your cluster been doing since you did that? Any heat issues or other problems?
u
nothing I've noticed, though I haven't exactly been running hard loads, and I've got fans in the case
in general my understanding is that you'd prefer to only apply silicone or other materials around the base of the components, ideally getting it into gaps, which means most of the surface should still be free to release heat into the air
so you may be able to apply it better than I did, though as I said, there's been no issues so far
r
Okay great, thanks. Now the only thing is Dow 744 is looking to be pretty hard to find. I don't suppose a more common one like 732 could work?
d
Which Turing Pi v2 board version do you have?
r
I have v2.4
Pi carrier boards are marked 1.0
d
Coil noise was fairly common with the combination of a v2.4 board, Pico PSU and CM4s. This was improved with the v2.5.2 board.
r
well too bad I only have 2.4 I guess.
u
I'd recommend getting in touch with a supplier and checking what they have that can be safely applied to the PCB. Alternatively, my research back 1½ years back suggests that a lot of people have had adequate results on similar issues with GPUs by using hot glue, but it comes down to the material properties of what you're applying in the end, and I can't say I'm an expert (which is why I asked my own local supplier back then)
r
okay thanks, I'll do some research
Okay my search has turned up one more option: https://www.amazon.com/Clear-Electronic-Grade-Silicone-Squeeze/dp/B0063U2RT8 ASI 388
Has anyone ever used this to deal with coil whine though?
If this also works it's much cheaper and readily available
This thread over at techpowerup.com talks about using Loctite 411 superglue also. Needs to be cyanoacrylate and not a blended type. https://www.techpowerup.com/forums/threads/how-to-quickly-easily-fix-coil-whine-coil-choke-noise.294015/
g
Have you been brave enough to try it?
That superglue thread seems promising
It looks like Amazon has a whole bunch of CYANOACRYLATE super glues. I wonder if any ole superglue of that type will work fine.
r
I had a correct type of superglue on hand so I've tried it. Going to wait until tomorrow for it to harden
g
I will excitedly await your update
r
Update: It's helped, but only a little.
I'm going to go back and try to identify where I did and did not get improvements and maybe try again
and I would say be extremely careful, I actually dripped a tiny bit of the glue into the end of Slot 3, but was able to use IPA and lot of patience to scrape it off and nothing broke 😥 😌
g
Ok. Good to hear any mistake was temporary. I ordered some of the correct type of glue from Amazon so I’ll give it a go sometime after it arrives
I’ll take any improvement I can get.
r
I went back and added some more glue to some of the components, before doing that I fashioned some card stock paper covers for the slots.
maybe there's a 3d printable DDR4 slot cover out there
g
Did you end up covering mostly the same areas as Usandru?
r
Yes but I went around the bases of the individual components where the glue would let me -- I didn't cover up entire areas of the board.
I'm pleased with the results
But for some reason NODE 3 always seems to make the most noise regardless of carrier board/CM4 module
One last observation that is especially anecdotal -- it seemed like a CM4 8GB revision 1.1 module in SLOT3 ran quieter than a 4GB rev 1.0
g
Thanks! My glue arrives soonish so I’ll have to give it a go
Well I’ve got my super glue and I’ve tried getting it close to the board near the orange things. It’s a bit hard to tell if I really notice a sound difference
The bright side is it still boots fine so I haven’t managed to do any damage
Ok. I think it actually did quiet things down some. I may have to go for a second round to improve it more.
I think a third application would quiet it enough i might even not hear it at all. Thanks @rayrayrayray you encouraged me to actually do something and it seems to have helped. That starbond super glue is hard to use though!!!!! Its so thin it runs every where. Require being very thoughtful / careful
r
Glad it worked out for you!
m
would any kind of super glue would work? got some gel gorilla glue but no idea if it's conduct electricity
s
My board also suffer from this issue. Can you tell us what changed (what solved this problem) between v2.4 and v2.5.2 from an hardware standpoint so we can solve the issue by ourselves?
r
I would be happy to pay for upgraded carrier boards if that can help also. I suppose we may end up with a new version to support CM5 also?
40 Views
PreviousNext
Powered by