Lenovo Coil Whine – Is it fixable? Or is it a reason to stay away from Lenovo?
I recently bought a Lenovo X1 Carbon. It makes an annoying high pitched sound. At first I thought it would be the fans and it will go away when I configure the dynamic fan speed.
But it turns out it is "Coil Whine" and that is common among Lenovo laptops:
https://www.google.com/search?q=coil+whine+lenovo
I cannot find any info on how to get rid of it.
So I will try to return the laptop. If I can't return it (Already removed Windows and replaced it with Linux), it would be a 100% wasted investment.
Since I otherwise liked it, I wonder what replacement to buy.
Are any Lenovo models safe from Coil Whine?
70 comments
[ 3.8 ms ] story [ 190 ms ] threadAnd what is real sleep mode? Suspend to disk?
These days we have "modern standby" instead, which keeps all things powered somewhat to allow for a quick wakeup by notifications for example. Just like your mobile phone. In theory, at least.
How do I check (from Linux) if the laptop can go into suspend to memory or not?
I think `cat /sys/power/mem_sleep` should tell you the desired modes; the active mode is selected and iirc changing one is done by writing the name of an option into the file. There are some related files next to `mem_sleep` in `/sys/power/`.
If you are moderately lucky, you may not have this _yet_, but you get it with a BIOS update.
But even if you have it, Linux might be hit and miss - sometimes it works, sometimes not, usually crippling some device like the touchpad after a wake-up.
You can try `cat /sys/power/mem_sleep` to see what states you have available. "deep" is what you want. "s2idle" and "shallow" are the ones that are just "low-power".
EDIT: actually, after a bit of web browsing it turns out that Intel 11th gen and above do not have S3 support anymore. So don't wait for BIOS support, it won't happen...
Would switching to an AMD laptop fix it?
Strange, that KDE still writes "Suspend to RAM" under the "Sleept" button when it cannot do that.
https://martinpoehlmann.com/posts/tech/2021-07-30-dell-s3-mo...
https://libreddit.privacydev.net/r/Dell/comments/h0r56s/gett...
And I don't want it to be able to wake up from this or do anything unless I press a button.
Not even that. Suspend to RAM is gone. It was perfectly fine, until someone had this great idea that PCs should behave like cellphones: just turn the screen off, meanwhile stay awake and connected. Hopefully it means doing the minimum necessary to receive and notify emails, but in practice this means the laptop will decide it's time to run the antivirus scan, do your scheduled backup and perform all kinds of maintenance right when your laptop is in your backpack and you're on your way to work. Microsoft calls it Modern Standby. I think there's also a degree of pushing on the manufacturers' side, as it's not purely an OS thing: sleep states must be enabled at the BIOS level for the OS to use them. Lenovo allows enabling the "legacy" sleep state, Dell does not.
See the previous shitstorm here: https://news.ycombinator.com/item?id=28646776
As far as I know, it's still no less of an issue to this day.
Specifically which CPU/RAM and display option?
I've just got a new X1 Carbon Gen10 with i7-1260P/32GB/WWAN/OLED screen, and it's one of the quietest laptop I ever used.
I did heard sometimes coil whine on my old Gen4 X1 Carbon, but not that it was super annoying.
I initially was spooked by the reports that 12gen Intel processors are not a good fit for an ultra-portable and may overheat and be too noisy, but that's not the case.
Did you had the same problem under Windows? Maybe it's some Linux-related problem?
In my case the only problems so far that webcam (that comes with OLED screen) and Fibocom modem are not yet supported under Linux ...
It is not Linux related, the sound is the same before I even booted Linux and go into the Bios settings for example. If you check the Google link in my post, you can see that it is a common problem with Lenovo.
I found this page that says applying superglue to the right component might fix it:
https://www.techpowerup.com/forums/threads/how-to-quickly-ea...
But I'm not sure if I want to do that to a laptop I just bought. Success seems kinda uncertain, but it certainly voids any right to return it.
Superglue or epoxy on an inductor is going to prevent the inductor from having sufficient cooling. As these are usually power converter inductors for various power supplies on the motherboard (GPU, CPU, etc.), they can get fairly warm. If you put superglue on it, it could overheat and literally melt ...
If the whine is a problem, you could try to have the laptop replaced but I wouldn't have much hope with that - it is working as intended (as designed ...).
Depending on where the coils are located, and if they are in some sort of airstream or not, I'd try that. But only in one piece, not several layered above each other.
Could result in dampening the whine, or eliminating it altogether, and better cooling. By pressing on the coil on one side, and some part of the inner casing on the other.
These days I have a nuc in an Akasa case to make it completely silent for my Linux desktop and an MBP for the road, I wish there was a good linux story for the laptop but I've just given up. It really is a shame as the tiling-wm story on macos is unusable.
I tried that. Not sure if it made any change at all. The noise is still too much to make the laptop useful.
Strangely, putting the laptop under load makes the coil whine go away. So it might be related to CPU power management somehow.
The fan speed that is also mentioned does not seem to have any role in this. The noise happens no matter if the fans run or not.
Probably be sub-harmonic oscillation. Under light loads the control loop can 'hunt' at a much lower frequency then the loop frequency. Combine that with an unlucky inductor and you get a high pitched whine.
It's an annoying problem because in susceptible designs it'll show up in a small percentage of some but not all production lots.
Take your laptop back.
When the load is low, a switching regulator can enter "discontinuous conduction" mode, where the inductor current periodically drops to zero. This is typically noisier than continuous conduction mode, where current is mostly constant with a small sawtooth ripple.
Dell batteries for example are trash, much higher failure rates.
I would like to read an explanation from professional electric engineer about how this sound may be fixed if it occurs on my device.
The noise is caused by one of the many inductors in the various power supplies (DC-DC converters) used in the machine mechanically vibrating at an audible frequency due to magnetostriction. This happens because of the power converter running at too low frequency - or hitting some kind of mechanical resonance mode that amplifies the otherwise inaudible noise. BTW, even ceramic capacitors can whine due to piezoelectric effect.
In a desktop machine you could replace the offending part (GPU, power supply, motherboard, etc.) for a different model but in a laptop you don't have choice as everything is soldered into the motherboard.
Even desoldering the offending inductor and replacing it usually won't solve the problem because it is due to the circuit design/properties themselves and not the inductor itself. So a new one will have the same issue - the whining inductor isn't defective!
I wouldn't recommend adding any sort of sound dampening because in a laptop this is very likely to compromise the already poor cooling. You will likely only make the otherwise benign (albeit annoying if you can hear it) problem into a very expensive to fix failure due to something overheating.
If you have a laptop with this issue and the noise bothers you, then replacing the laptop is your only realistic option, really. Or wait until you get older and will naturally lose the ability to hear higher frequencies (past 40 years of age you will be hard pressed to hear anything above 15kHz anymore ...).
Either way, before attempting any sort of remediation, do make sure the noise is actually a "coil whine" and not e.g. the speakers emitting it because of some unused mixer input that is set to high gain and is picking up and amplifying internal machine noise.
And yeah, good luck with trying to return a machine for this reason - the computer is not defective, this is a normal behavior (in the sense it works as designed). So it is unlikely the seller will recognize it as a warranty issue of any sort.
I have read that opening the laptop and putting superglue on the right parts might reduce the noise.
These wonderful "fixes" (along with people "reflowing" stuff using heat guns) are a great way how to turn an annoyance into an expensive to fix failure.
This makes the noise even louder:
This brings it back to the original, already too annoying, level: So it seems somehow related to the CPUs.I use powersave capping the max freq unless I have something big to compile etc (just to help make sure the fans won't kick in, in my case, but it might help)
Just write whatever frequency you want to /sys/devices/system/cpu/cpu[0-9]*/cpufreq/scaling_max_freq and scaling_min_freq
The lowest I can set it to seems to be 40000. I can still hear noise, depending on what the laptop does.
As simple command to create noise is:
It makes noise even when the terminal is not visible.I guess I would lose about 90% of performance if I permantently set scaling_max_freq to 40000?
I normally scale down to about 1GHz without much visible performance impact except for compiling, and that is enough for fan to not kick in randomly everytime a web page has some somewhat heavy js or whatever it is I'm doing, but obviously you're looking at a different beast.
The coil whine was (relatively) less of a nuisance than the X1's fans, though, which were ridiculously noisy until I set the thing to 80% CPU speed.
So yes, Lenovo, if you're reading this, your customers have to kneecap your hardware to use it. Way to go proving value here.
I mainly have heard it happening to desktop GPUs, both nvidia and amd.
I have a Thinkpad T490s. No coil whine!
>Are any Lenovo models safe from Coil Whine?
Yes :)
I suspect most users either do not hear it or do not find it that annoying. Maybe there will be more focus on it once fanless (and otherwise totally quiet) laptops and desktops will be more common, so that coil whine turns into an issue that UX designers care about.
My advice, file a claim under warranty and try your luck, because even if they accept it as a warranty replacement and swap out the laptop, the next one might very well suffer from the same thing (as you pointed out, it seems a common thing with lenovo)
Does it support S3 sleep state?
One downside I see is that it seems to be 130g heavier than the X1 Carbon series.
I don't bear an absolute grudge against Intel, but they really need to get their act together when it comes to laptop chips, the AMD chip is a dream compared to the Intel chips I've used in laptops for the past few years (my desktop 12700k is just fine, though).
130g is inconsequential, it feels very light and well distributed. After all the struggles, and a financial hit selling the X1 two years earlier than planned because it was just unusable, I had to determine, for now, the compromises of the Intel X1 series are not worth it, though they would probably be fine with an AMD chip. I do miss a touch screen and very occasionally convertible modes, though.
The way I read it, s2idle leaves the CPU running. So it is up to the OS if it wants to use it or not.
https://www.kernel.org/doc/Documentation/power/states.txt
Too bad. I was hoping AMD still supported it.
My Lenovo X1 was fine. They might swap it if it’s new.
I have, however, heard it plenty from laptop chargers, especially of the Dell variety, and while I guess that's somewhat expected given the transformers converting from AC to DC, it still skeeves me out a bit to hear noises from the one thing in between my laptop and 120VAC.
I wonder if it'd be possible to use destructive interference / noise cancelling to "cancel out" coil whine?
In any case, it seems excessive to call a laptop with coil whine a 100% wasted investment. Even if you can't return it, you could always stick it in a cabinet or something and use it as a desktop or server or media PC. Or hell, surely you could find a buyer who's too deaf to notice ;)
Try to return it without explaining why. Otherwise they will try to fix the problem by sending you another laptop which has a high likelihood of having the same issue. Best is to just get your money back and avoid doing business with them in the future.