The ULTIMATE guide to building a silent PC

Introduction

Almost all of us would like to have a PC that does its work as quietly as possible, if not inaudibly. This tutorial provides you with everything you need to know to understand what makes a computer loud. And of course, what you can do to make your PC at least a little bit quieter.

Of course, this complex and extensive topic can't be fully covered in a few sentences. But don't worry; you don't have to go through the whole set of pages to get a good overview. I would recommend that you first skim the 'General' section to learn what the major noise sources are, and then look specifically at the subsections that deal with the sources that in your specific case are the most important ones.

Contents

Why is my PC so noisy?

To clarify this question, we first need some basic understanding of acoustics. Don't worry, I won't go too deep into this topic at this point and if I am completely honest, then my last lecture on 'acoustics' was when I was still at university and this was a good 20 years ago. And I also want to make this as simple to digest (and apply) as possible.
One thing is important in this context is the fact that the volume level of several sound sources cannot just be added together (in the sense of two sources with the same individual volume are NOT equal to the a sound source with a volume that is twice as loud). If you want to know exactly: for something to be perceived as twice as loud, the level would have to increase by 10 dB, which would correspond to an increase in power by a factor of 16. But physics and (psycho)acoustics aside; that is not what this article is about. We just want a quiet PC!

So, based on what I have just said you might now think: "Well, in that case a second sound source won't make that much of a difference." That's kind of true, of course, but if our goal is to build the quietest possible PC, then unfortunately that also means at the same time that the overall system won't get much quieter if you only get one of the noise sources quieter when you have two equally loud sources. In the end, unfortunately, you have to identify all noise sources and eliminate all of them if you want to achieve an optimal result. If this effort is too much for you, however, then by eliminating the loudest source you have the most leverage to get the overall system as quiet as then still possible.

One more thing to know and keep in mind about this topic is that the same sound pressure levels of different frequencies also perceived to be 'differently loud'. And again, I'll refrain from going into details about dB(A)-evaluation of sound pressure levels (which strictly speaking is only valid for sinus tones anyway) and shorten this to the important essence: It also depends on the frequency and 1000 Hz are at the same sound pressure level sounds to the human ear louder than a sound with 250 Hz. In addition, the pure 'loudness' is not the only criterion, but also something like the 'annoyance' plays a role and there is unfortunately no good and objective parameter to evaluate that. But what everybody knows from his own experience is that our brain is quite good at filtering out uniform sounds after a short period of time, so much so as that we sometimes no longer consciously perceive them. It's a different matter with variable sound sources though, especially if they are constantly changing in volume and/or frequency and can literally drive us crazy.

But our goal is to have a PC that is as quiet as possible, and we now know that the loudest source determines the result. And that's what we have to tackle! Fortunately you almost never need expensive measuring equipment to identify this sound source. Most of the time, our ears are sufficient to find out where the noise we want to stop is coming from. And what's even better: in order to see what effect our optimization measures will have, you can you can temporarily 'silence' almost all noise sources in the PC, at least for a short time. Fans can be stopped, hard disks — apart from a boot drive, which should hopefully already be noiseless in the form of an SSD — can also either put to sleep mode or simply disconnected from the power supply (of course you should turn the PC off before you do this!).

So much for the basics. In the following sections I will talk about all the components that you will find in your PC and that can (potentially) cause annoying noises. The remaining sections will then focus on individual components. More exactly, I will explain in each case where the annoying noise originate from and give solutions, how to reduce or eliminate those noises as much as possible. At the end, you'll find some basic considerations on the topic of silent computer cases. Finally, I'll discuss the possibility of using the distance between the PC and your ear so that the annoying noises of the former no longer reach the latter.

Where does the noise in my PC come from?

First you have to know that noise strictly requires that something vibrates and that the vibrating surface comes into contact with the ambient air, so that this movement can be transferred to the air molecules. Because, ultimately, these annoying noises are nothing else than sound waves in the audible frequency range (depending on the age of the listener between 50 Hz to 10.000 Hz or even up to 20.000 Hz). So, clearly, without oscillations there can be no annoying noises. But I think most of you already knew that and also the fact that noises disturb us mainly when they hit our ears sounds logical and makes sense. But from this follows at the same time that we can live with the most annoying noises, if they originate inside a computer case — and stay there. The thing about 'staying there' is unfortunately the problem, because where the noise stays, so unfortunately does the heat, and that can only work for a very short time, until modern gaming PC has to throttle its performance, shuts down completely or even worse: gets damaged. Nevertheless, a computer case with good sound insulation is a good way to significantly reduce the noise level. And that's why there's a section at the end on silent PC case.

The rest of this article will now give you an overview of all the components that make those annoying noise in your PC. You might find some things that you can simply ignore, because because I might talk about things that either don't make any noise in your PC or is hardware that is not even installed in your PC. But of course I don't want to go too far and have left out some exotic things. Either, because they are things that almost nobody uses or things that are almost extinct. If you still use floppy drives, ZIP drives or DAT tape drives and you are annoyed by the noise they make, you should consider getting rid of those and replace them with better — and more importantly: silent — alternatives.

General

Despite the arrival of alternative, faster and also noiseless storage media, such as SSD drives, almost every PC still contains at least one classic hard drive. And unfortunately, in addition to the case fans, the Hard disks are one of the two top sources of noise of a PC, the other one being the multitude of fans scattered around the case and attached to many components. Contrary to what many people might think, it's not the operating noise emitted by the hard disk itself that's important here, but most of the annoying hard disk noise originates somewhere else. Why this is the case — and what how to get them under control — that I want I aim to explain in this section in more detail.

The Cause

The operating noise of a hard disk drive results in large part from the friction created when the spindle that holds the platter(s) of the drive rotate in its bearings, and in other part from the noise created when the actuator arm moves the read/write head(s) to position it in oder for it to be able to read or write the data. The hard disk itself emits a part of the noise via its casing into the PC's interior; however, as already indicated, a large part usually arises at a much more distant location. Since a hard disk is connected to the case of your PC, its vibrations are inevitably transferred to the case and are transmitted by structure-borne noise inside the case. This also creates vibration in the side walls of the case, which in turn act like the membrane of a speaker and emit the noise to the outside. Since these noises are actually outside the PC, an insulation inside the case itself unfortunately has little to no effect when it comes to reduction of hard drive noise. There are still things that you can do to a case and which are described in the section about silent PC cases.

The Remedy

The ideal and sadly for many of us also unaffordable option is, of course, to forgo the use of hard drives altogether and stick to noiseless alternatives in the form of SSDs. Unfortunately, the prices for SSD are still significantly higher than those of a hard disk of the same capacity. Furthermore, SSDs in the size of one of the larger hard drives are (currently) not yet available, at least in the consumer segment. So, changing all HDD to SSD might be a possibility in the future to significantly reduce the noise emission of our PCs, but for now we have to look for a solution how to achieve this with our still needed hard disks.

As just described, we have to look at two sources of noise when it comes to hard drives. The best way to get a handle on the operating noise generated by the hard drive itself is to buy a suitably quiet hard drive. If the hard disk is only rarely used, for instance only for an PC internal backup — whereby this is strictly speaking is no 'backup' solition at all!; but to go into this in more detail would go beyond the scope of this article — then you can let this disk drive enter sleep/standby mode when it is not in use using the windows energy settings. The disadvantage is that the disk has to wake up before each access (which can take quite a long time and a delay of 10 seconds is to be expected). And even if you don't want to access the disk, a hibernated disk can lead to unpleasant delays, e.g. if an application wants to display all drives when opening a dialog and the necessary accesses lead to the disk waking up. The same applies to accesses to the S.M.A.R.T. attributes, e.g. for reading the hard disk temperature, because also for this action the hard disk must wake up again and be set into rotation. All in all, you have to decide how important the operating noise is to you and often, with modern hard disks, the direct operating noise is less disturbing in a well insulated case. But of course, everyone has to evaluate that for herself.

The second and more important noise source of a hard disk is the emission of the hard disk vibrations over the side walls of the case. Although it is possible to reduce this inside the enclosure itself if necessary by using heavy insulating mats made of bitumen, the best way to go here is to prevent the transmission of vibrations from the hard disk to the case as far as possible. For this purpose, there are appropriate mounts where the hard disk is either decoupled from the case. In some cases, rubber buffers are used for this puprose. In other mounting variants, the hard disk is clamped in rubber cords and only held by them. These rubber bands themselves are stretched into a frame, which is then connected to the housing. Unfortunately, the decoupling by rubber buffers has proven to be less efficient, because if they are too thin or too tight then the vibrations are still continue to be transmitted well to the PC case. Also, with all these changes to how the hard drives are fixed, one thing is very important to mention: the purpose of bolting them to the case is not just to fix the hard drives in one place, but it also dissipates some of the heat to the case via the connection. No matter which decoupling is chosen; if there is no longer a good contact with the case, then ideally, less vibrations can get into the case, but in any case, the heat transport is also effectively prevented and it is advisable in any case to provide for a good cooling and to keep the operating temperature of the hard disks in the eye to keep. Programs such as Argus Monitor can be used to monitor the temperatures of the hard disks, generate a warning if temperature limits are exceeded or can even control chassis fans depending on the temperature of the hard drive.

General

In this section, I will address one of the two main causes of annoying noise that our PC makes. The other is hard drive noise, to which I have dedicated a separate section. Since no PC can survive for long without sufficient cooling, and completely passively cooled PCs are probably pure exotics that should only make up a tiny fraction of all PCs, it's almost certain that your PC has a couple of of fans installed and that they will also be responsible for a large portion of the total noise, which your PC will give off. From that it follows naturally that getting the fans in your PC to operate more silently has a large potential to make your PC MUCH more silent or — ideally — turn your PC at least in certain scenarios completely inaudible.

The Cause

Fan noise is caused both by friction within the bearings and directly by excitation of the air molecules by the turbulence on the rotor blades. In addition, case fans can also introduce vibrations directly into the housing. Other than for hard drives, for which this component is the major contributor of noise, the excitation of the PC case by the fans plays a rather subordinate role. The task of fans in your PC is to guarantee the airflow necessary for cooling the installed components. The faster the fans rotate, the greater the airflow. Unfortunately, this is not only true for the airflow, but also for the fan noise, which also gets louder with increased fan speed. Therefore it is necessary to find a compromise between 'acceptably quiet' and 'still fast enough for sufficient cooling'. Since larger fans have to run significantly slower to produce the same airflow, you should make sure to purchase the largest possible fans that will fit in your existing case. Or, if even the purchase of a new case is pending, let the size of the fans you can mount into it also play a role in your purchasing decision and select a case in which you can install a sufficient amount of large fans.
A side note in terms of psycho-acoustics: the fan noise caused by changing fan speeds is perceived by us humans as much more annoying than a constant, but perhaps slightly louder louder noise of faster rotating fans.

The Remedy

Before any other measure you should first make sure that the fans can actually do their job. Usually overt time there accumulates quite a lot of dust inside your PC, especially in PCs of pet owners, to at least hinder smaller fans in their work. Occasionally, it also happens with older fans that they emit a humming noise. This is the result of a defective bearing and the only real remedy is to replace the affected fan. However, a change of the installation position can also help to operate the fan without this humming for a while. (so e.g. with a fan at the housing back by exchange with one mounted on the top).

As already described several times, the annoying fan noise is proportional to the fan speed. This means that by reducing fan speed, you have the greatest potential to reduce loud operating noise from an existing system and turn it into a quiet PC. Of course, it is important to keep the actual task of the fans in mind. They have to provide the airflow necessary for cooling the components and therefore when reducing fan speeds to reduce fan noise, you also at the same time have to monitor the temperatures of all components. Ideal are programs such as Argus Monitor, which can perform both tasks simultaneously, i.e. control the speeds of all fans, but also display the temperatures and also issue a warning if the temperature of a component reaches a critical threshold value. Of course, the optimal solution is if you can can directly control the fans depending on the temperature of the components.

A basic option for fan control is usually available directly in the BIOS of the motherboard. The disadvantage of this solution is that usually only very basic control options are available and that one has to restart the PC in order to change the control parameters. This solution can also only access sensors installed directly on the motherboard as well as in most cases also the CPU temperature. A fan control depending on the graphics card temperature or the temperature of hard drives is not possible when using your BIOS for fan control.

In addition to control via BIOS, various software programs are available that allow fan control to be more flexible and dependent on other temperature sources. The most well-known example is the freeware SpeedFan, but its development was unfortunately discontinued back in 2015. Also, the configuration is overly complicated, but if you are searching for a free solution for an older hardware, SpeedFan might be worth a look.

Another and also free option are the software products that are included with almost every motherboard. They also allow the control of all fans directly connected to the motherboard, but are usually rather confusing and inconvenient to configure and also very limited in the selection of temperature sources. Like the BIOS, they also only offer the CPU temperature as a temperature source, as well as the sensors installed on the motherboard.

The most flexible and comprehensive option for fan control via software is offered by Argus Monitor. Now, for the sake of honesty I have to confess that — as one of the two authors developing Argus Monitor in their spare time as a hobby project — I am not completely unbiased in this assessment. But objectively, there are so many points that speak for the statement that Argus Monitor is the best program for controlling PC fans, that I can say so with a clear conscience and a straight face. And when I write about fans, this includes all fans and not just case fans, but also CPU fans, fans of various AIO water cooling systems and those of graphics cards.
With Argus Monitor, you can either simply control the fan speed dependent on the temperature of any component (just select the one for which the fan in question provided the airflow for cooling. It is there fore possible, for example, to control of a case fan by using the temperature of the graphics card or the highest hard disk temperature as an input. This is something that is neither possible when using the BIOS for fan control, nor if you opt for the software that is included with the various motherboards to do this job. However, if you want to exploit the full potential, then you can also make the fan speed dependent on all the temperature sensors present in your PC. More precisely, the control of the speed can be done by an individual characteristic curve. This can then be made, if desired, dependent on e.g. the CPU temperature, but just as well on the GPU temperature, the temperatures of hard disks, the water temperature of AIO coolers or the temperatures that are various temperature sensors installed on the motherboard. Furthermore, it is possible to calculate own, synthetic temperature values from the existing temperature sensors. You can use functions such as mean value, temporal averaging, maximum, difference (e.g. from water temperature of an AIO cooling solution and ambient temperature) and more to generate an individual artificial temperature value to be used as input for the fan controller that is tailored to your own system. On top of that, other measures like smoothing or hysteresis can be used to ensure that small, short-term fluctuations in the temperatures used for control do not result in annoying speed changes.
Furthermore, with Argus Monitor it is possible to operate several controllers in parallel for each individual fan and then determine the necessary speed of a fan from the maximum value of these controllers.
In addition, the available fan profiles offer the possibility to define your own controller settings for different use cased, which can then be easily switched between at any time.

General

There might not be a graphics card in every PC nowadays, as CPUs with integrated graphics units can be found for the less demanding user (so-called APUs — actually a marketing term from AMD, Intel calls them 'CPU with integrated HD graphics'), but as soon as the demands on the graphics performance are are higher, there is no way around a dedicated graphics card. With these, there are again two different noise sources, whereby I have dedicated a separate section to one of them and will in this one only focus on the other one, which is of much more importance to most users.

The Cause

With graphics cards, as mentioned, there are potentially two sources of annoying noise, one of which fortunately only affects a small portion of graphics cards. This is the noise generated inside the coils of the coltage converters and manifest in the form of so called coil whine, about I have written in a separate section. The other is the more obvious problem, and it affects almost every modern graphics card: the fan noise caused by the graphics card fan(s). And since, with a few exceptions, no graphics card can do without at least one — more often two or three of these — of these fans, this source of noise can also be found in almost every PC.

The Remedy

In this section I limit myself to the possibilities of intervention in relation to the graphics card fans, because for the more exotic topic of VRM coil whine, there is a separate section in which I describe the somewhat more complex causes and ways of reducing this problem. So let's focus on the fan noise, and in that respect the same that was just said before about other fans applies as well: the smaller the fans are and the faster they turn, the louder. Now, one has no influence on the type and size of the fans installed on a graphics card, i.e. in this regard, you have to make the decision before buying and then at least live with the installed hardware. There are also different cooling concepts for graphics cards, whereby some suck in the air in the case via a radial fan and and then transport it out of the case through the cooling fins of the GPU to the back via the slot bracket. This is almost always the noisier option. Another variant is to suck in the air inside the case via axial fans and then then release the warm exhaust air back into the case, leaving its final 'disposal' to the environment outside the PC case to the case fans. This has the disadvantage that it gets even warmer inside the case, which also affects the cooling of the other components. Acoustically however, this variant has the advantage that the waste heat can be transported much more quietly into the environment via the much larger case fans. But once you are settled with one specific GPU, one has — apart from the choice of the model when buying — only one way to influence and reduce the noise a fan of a graphics card will emit.

If one wants to get an existing PC to operate more quietly without exchaning the GPU altogether, then the only thing you can do is to control the GPU's fan speed. The good news here is that many modern graphics cards usually completely turn off their fans in desktop mode and therefore in most cases perform their service silently. Luckily, most the fan control characteristics of modern GPUs are usually good enough that manual intervention is rarely required. Only in case you want to get an optimal result and control the GPU fan speed yourself, you find programs such as Argus Monitor that will allow for an individual fan control curve to be used.

General

Liquid cooling solutions have been making an increasing appearance in the cases of many PC enthusiasts lately. This is probably mainly due to the fact that all-in-one (AIO) liquid coolers are almost as easy to install as conventional air coolers. An AIO liquid cooler is a sealed system that comes already pre-filled at the time of purchase, where a heat sink with an integrated pump is connected to a radiator with two hoses. So, no cooling liquid has to be filled in, but the heat sink has only to be installed onto the CPU (or GPU) and the radiator in the PC case. AIO liquid cooling solutions take up less space in the direct vicinity of the CPU and thus leave room for other components. In addition, the tidier interior is especially popular among PC owners who run cases with a transparent side panel. A disadvantage in this context may be that the air flow that is normally produced by the conventional fans on the CPU coolers is missing, which one would normally be able to use in conventionally cooled systems to cool the voltage regulators or RAM modules with.
In this section, I will mainly focus on AIO liquid cooling systems, although most of the explanations will in principle also remain valid for models with an external pump.

The Cause

For normal liquid PC coolers, we also have to consider several independent causes for the emission of noise. First, as with the graphics card and the case itself, there are the fans that provide the necessary air flow for better heat exchange at the radiator. The points of view on the subject of fan noise do not differ from those that I have already described in detail in the section dedicated to this subject and I will therefore not go into it again here. In addition to the noise caused by the fans, there are two other sources of noise for liquid coolers. On the one hand, there is of course the pump noise, but even in 'high flow' configurations (i.e. where the pump is operated at high speed) the pump usually operates relatively silent. But there is a second source for noise in liquid cooling systems, which almost always can be eliminated completely with the right setup: the gurgling and bubbling of the coolant or the howling of the pump. So I will first briefly discuss the normal operating noises of the pump and then turn to the second cause in the rest of this part of the guide and help you to get these noises under control.

The Remedy

As just mentioned, the noise of the pump itself is usually of rather secondary importance. Even if you can occasionally read on the Internet that cooling with a lower pump speed works better, because then the water has 'more time to cool the CPU' due to the slower flow speed, this statement is just plain wrong (and almost painfully so). The cooling performance of a water cooling system is better, the higher the flow velocity of the cooling medium is. As with fans, the same applies here: higher pump speed, better cooling performance, but also more noise. On the other hand, liquid cooling systems, if properly installed, are already very efficient cooling devices even at low pump speeds. So, except for overclockers, a medium or low speed should be sufficient to keep a CPU or graphics card within the permissible temperature range. But of course it never hurts to measure and keep a look at the CPU temperature. Or maybe even control the pump speed dependent on the CPU temperature with a program like Argus Monitor. But as just mentioned, for the pump speed and regarding noise this is rarely really necessary.

The second cause is much more annoying, but fortunately you can usually eliminate this kind of noise completely. Unfortunately, there are indeed many installations of AIO liquid coolers, which are characterized by a more or less noisy bubbling or gurgling of the coolant which might in extreme cases even be accompanied by howling pump noises. Probably by then your first thought might be: "OK, that must be like liquid cooling systems sound like." or even: "Damn, the pump of my AIO is broken. I guess I'll have to replace that unit." And more often than not, that's not the case, but they cause by mistakes that were made during installation and which can be fixed rather easily. So, what exactly is 'broken' here and how do you get the AIO as quiet as it should be.

The cause of this type of noise is almost always the air that is inevitably present in any AIO. Even right after purchase, 2%-10% air in the system is normal, and over time, this decreases percentage increases even more due to permeation. And this trapped air will always rise and gather at the highest point in the cooling loop. The gurgling and bubbling will occur when this air is entrained by the flow of coolant and then rises back up through the radiator. The howling of the pump in turn indicates that there is an air bubble inside the pump, causing its speed to fluctuate. Some pump models also use the coolant to lubricate the pump itself. If this is the case, not only will the pump whine annoy you in the short term, but soon you will also have to invest money in a new cooling solution. With other models, this is not a problem with regard to the durability of the pump, but the pump noise is still annoying and of course the cooling performance of such a system is extremely limited, meaning that the CPU will quickly reach Tjmax and either the system shuts down altogether or at least clocks down, which will be noticeable in a performance drop. Both then lead many users to the assumption that the AIO cooling system would be defective, whereas the cause here is actually 'just' an incorrect installation.

So what do you have to consider when installing an AIO water cooling system? So, what now is the most common mistake and of course above all: How do you install an AIO liquid cooling system correctly?
As already mentioned, you must ensure that the pump is not at the highest point in the loop. In general, most common mistake when installing an AIO cooling system is that the radiator is installed at the bottom of the case. This placement is to be avoided at all costs, because that way the pump — which is located in the CPU heat sink — will always be at the highest point and it is, so to speak, ensured that the air in the cooling system will find itself exactly at this point. A bad cooling performance and in almost every case also pump noise are guaranteed in this constellation.
As far as themals are concerned, then the most favorable installation of the radiator would be at the front of the case at the intake side of the case, but also the installation at the outlet (usually in the top of the case) is possible. When installing the radiator in the front of the case, the hoses of the cooling system should be at the bottom end of the radiator, if possible. Only if that is not possible — e.g. due to too short hose lengths — one can install the radiator also in such a way that the hoses are on the top end of the radiator, but you must then ensure that the upper edge of the radiator lies still clearly above the pump.

General

A power supply is one of the components that can be found in every PC. And also one that almost always makes more or less noise, at least if your PC doesn't use a passive or semi-passive power supply. So, if you are willing to spend some more money, then you can indeed get a device that — unlike its brothers, which transport the waste heat out with one or two fans — at least often (with semi-passive) or always (with passive ones) do their job in a silent fashion. Unfortunately, however, power supplies are also a place where you are very limited in what you can improve if you find yourself with a noisy power supply in your existing PC. That also means that with power supplies it is especially important to pay attention at the time of purchase, to select of the correct power supply unit, with the required power rating, efficiency but also a low noise level.

The Cause

The main source of noise in PC power supplies are usually the installed fans that provide the air flow to transport the heat from inside the power supply to the outside. And for that specific topic I have written in a separate section about fan noise. What is important in general is that all the heat a power supply produces comes from the power dissipation that occurs during voltage conversion. Therefore, the more efficiently the power supply is, the less heat is generated and the slower (read: quieter) a power supply fan can turn. This even goes so far that some power supplies stop their fan(s) completely at low loads and thus almost can perform their service without any of the annoying fan noise. But if they do rotate, then the smaller the fan and the faster it spins, the louder it will be. And unfortunately, what is usually an option to reduce fan noise in the case, CPU and GPU fans, namely software fan control, is rather rarely the case for power supplies.

The Remedy

But this also means: with power supplies, the options to redure fan noise are unfortunately very limited. At least as far das existing systems are concerned. If a fan control option is provided — which is rather rarely the case is —, then one can lower the noise level somewhat with a suitable fan curve.
Argus Monitor for example supports the regulation of the fans of the current Corsair AX power supplies. But if the power supply is really the loudest source of noise in the PC, then usually a replacement by a new, quieter model is a real long term solution. As a general rule, it can only be said again, that — assuming the same performance category — power supplies that have a higher efficiency are usually quieter because there is less waste heat produced by the inefficiencies during voltage conversion, which a fan has to transport away. But regardless of that, it's strongly recommended to consider the topic of noise emission in addition to overall performance and cable management when buying a power supply.

General

Now, when I write about voltage regulators here, one or the other may now interject: »Didn't you say that it has to rotate, vibrate, or otherwise move a bit for annoying noise to arise? Nothing moves inside a voltage regulator!. So such a part CAN'T emit any annoying noises by itself!« Yes, I had written that and at first glance (and perhaps also on the second and on sleepy Sunday mornings possibly also on a third) it probably is the case all parts of such a voltage regulator are soldered into forced rigidity onto a circuit board to endure its dreary existence. But if you look at this a bit more closely, then something moves there nevertheless. And these somethings can — at least if one has at least a bit of bad luck — be really quite annoying.

But first things first! What is it that is moving? Where are those noises coming from? And most importantly: How do you turn it off?

At this point, a short insertion on the topic of voltage regulators. If you are less interested in understanding the cause and origin of the noise, but only in how to fix it, you can skip the next section and turn to possible possible 'countermeasures' further down the page. For all others, let me allow to digress for a bit and try to bring the matter a bit closer. I will keep it on the level of a technically interested layman. OK, let's get started.

The Cause

In our modern computers, there are quite a few of these voltage regulators installed, whose job it is to convert the voltage of the switching power supply (e.g., from the approximate 12V DC voltage), to the working voltage(s) (1.2V, 3.3V, 5V, …) needed for the individual components on a circuit board, e.g. your GPU. Most of these devices actually perform their work silently, and that is when the currents flowing through them are relatively low. Unfortunately, this isn't always the case, because some PC components require enormous currents and therefore the best known and unfortunately also the loudest representatives are found on modern high-performance graphics cards. The loudest representatives of those VRMs are found on modern high-performance graphics cards, where these noisy things are often found in large numbers.

Somewhat simplified, the core of such a voltage converter consists of an coil, a capacitor, an electrical switch (MOSFET) and a actuator for that switch (a PWM controller on the gate pin of the MOSFET), which then lets current flow to, and — after the build-up of the induced magnetic field — through the coil. Now, such a current is always accompanied by a magnetic field. And in a magnetic field, a force acts on the wires a current flows through: the so-called Lorentz force. But this force is not constant, but of course is changing with the same rhythm with which the magnetic field in the coil builds up and dissipates. And unfortunately, this often happens in a frequency range that is perceptible by the human ear, and then manifests itself in a more or less annoying whine, the so-called coil whine. What you hear exactly are the coil windings excited by this Lorentz force, which scrape and rub against each other (and also against the coil housing). Depending on the load of the voltage transformer, the whine can have a different amplitude (which determines the volume level of the noise) and whistles in different pitches (depending on the frequency of the PWM signal).

The Remedy

Now that we know how the noise is generated, we can consider how to get it under control most efficiently.

Of course, the simplest (but in most instances probably unsuitable) solution is to eliminate the flow of current through the coil ⇒ turn your computer off and spend your time on the patio, in the park or garden. No current, no Lorentz force, and no whining. This works, but who can (and wants) to spend that much time outdoors and away from his PC? So, no, there must be another way.

So, what do we really have to do if our PC has a graphics card installed that spoils our game sessions with that annoying coil whine? Fortunately, there are a few things that we can do in order to address the problem and if not completely eliminate so at least strongly to reduce the annoying noise. Often, the coil whine is loudest when the graphics card renders a very high number of frames. Unless you have an uber-mega-monster PC — or are playing a game with no graphics requirements whatsoever — then this occurs most often in menu screens, where the graphics card can easily render several hundred frames per second. As long as this FPS number lies well over that of the monitor, then the only 'benefit' of these additional frames is an increased power consumption. And with that also a clearly perceptible coil whine. The obvious remedy here is therefore to limit the number of frames to what is necessary and reasonable depending on the intended use. How many FPS you should aim for depends on your specific monitor model, its refresh rate and the supported technologies (G-Sync/FreeSync). For instance: to drive a 144Hz G-Sync monitor with more than 144 Hz rarely makes sense and therefore a good possibility to get rid of coil whine is to use a FPS limiter to cap the number of frames to a value below the monitor refresh rate — in the mentioned example e.g. to 140 or 141 Hz. Ideally, the FPS limiter that you should use is the one that the respective game already comes with. Only if this is not possible, you can also fall back on an external limiter, e.g. by using external tools like, in the case of an AMD graphics cards, the AMD 'Frame Rate Target Control' tool, the tool included with MSI Afterburner, namely 'Rivatuner Statistics Server' or by setting the corresponding entry in the Nvidia control panel. When using this approach, you have to make sure to never! use two limiters simultaneously! If you can set a limit in the game, then only use this and disable all other limiters.

With an FPS limit alone, you can in most cases already significantly reduce the problem of coil whine. But another approach, and one that is not quite as obvious, is to reduce the noise-causing friction of the coil windings directly. Since we cannot reduce the current and thus the force with which the coil windings are excited, and since the frequency of the PWM signal is determined by the design and cannot be changed, the only thing that we can do is to dampen the windings themselves. At this point, however, let me start with a warning: What I describe on the rest of this section is something that has proven itself well in practice, but where I must must point out that this kind of tampering can damage your hardware and I can not accept any liability for it.

So, what is it about and what else can be done about this coil whine? As mentioned several times, the noise is caused by the coils' windings rubbing against each other and against the coil housing, and that's exactly where you can reduce the noise, namely by making it more difficult for the coil windings to move by giving them no or less space to do so. At this point I want to strongly advise against experimenting with hot glue, because it does little or nothing (apart from destroying your graphics card). The means you can use to restrict a coil winding's freedom of movement are other coil windings. And you can achieve this by making them take up more space. As it turns out, the coil whine is a temperature dependent phenomenon. The warmer a coil is, the more space the windings need and the less they can move and rub against each other. Now, the only means to influence the temperature of these coils is to reduce their cooling specifically and moderately so that they become somewhat warmer, take up more room and and thus get quieter. This works quite well in practice, but of course one must not overdo it and also here it depends on the respective graphics card, how to best achieve this modification, because there are of course components in the immediate vicinity of these coils that would possibly resent a reduction in cooling and you therefore have to check in each individual case if and how this approach can be implemented.

Now that I have taken a closer look at all components of an PC one by one and have given hints on how to get a handle on the noise each one makes individually, let's look at the one thing that has the potential to affect the noise output of the overall system in general: the case. If you already own a PC and are looking for ways to modify the case in order for it to let less noise leave the case overall, then the options are unfortunately very limited. At least under the assumption that you have already used your case with all covers mounted securely in place. All things considered, there is one rule that of course always applies: wherever air comes in, noise can come out. And where and how this noise is being generated we have already illuminated in detail in all the other sections before.

The main task of a PC case is to accommodate all the components of your PC and also case fans (or those on the radiator of a liquid cooling system) to generate the required air flow to prevent a heat death of all your expensive hardware. This airflow is usually directed that way: Intake at the front/bottom, Exhaust at the back or top. The alignment of the fans ensures that this works as desired. If you are not sure in which direction your fans move the air, then there is usually also an arrow on the side of their housings, which indicates the direction of the air flow.

A big mistake is to assume that you can get a PC to be quietest if you close off the front and let the fans on the back or top — that is, as far away from the user's ears as possible — just pull the 'hot air' out of the case. But of course, they can only suck out what comes in somewhere else. And closing off the front only ensures that the fans are forced to spend a lot of effort (read: high speed/loud noise) to pull the air out of all the gaps of your PC case. Therefore: if we want to pull hot air out of the case, then we also have to make sure that cold ambient air gets into the enclosure 'at the other end' and can flow past all the components that need to be cooled. For that reason, a relatively open case with good airflow will usually be significantly quieter for the same cooling performance as one that is well insulated, but in which the fans have to run at high speed to transport the hot air out of the housing at all.

But you should still not neglect the insulation of the case itself. The side walls made from thin sheet metal that are frequently used with inexpensive housings, can be considered themselves as kind of membranes, which will then be able vibrate turn your PC case into a loudspeaker. In addition, the natural frequencies of uninsulated case panels are often in a range where they are excited by the vibrations generated inside. Therefore, when buying a new case, one should pay attention to get one with thicker material for the side panels.
To improve the 'membrane situation' there are damping mats available, but from experience it is clear that light, cheap dampings are surprisingly ineffective, because they 'swallow' some some sound from the sound generated inside, but at the same time have little to no impact when it comes to the problem of the membrane effect present with too thin side walls. Even then, sound waves that are introduced into the housing (mostly from hard drives) are still radiated to the outside. For noise insulation is to be effective, it is best to use bitumen mats. These are heavy and can change the natural frequencies of the enclosure by enough (shifting them to lower frequencies) in order for them to no longer get excited as much by the spectrum of vibrations. A clearly better and more important measure is still to eliminate the noise problem at the source and to prevent the excitation by the installed hard disks. What you should consider in that regard I have described in a separate the section on the topic of hard disks.

When it comes to case fans, then the same applies here as it was already been discussed in the sections about water cooling and (case) fans above: ideally, select case fans that are as large as possible. Those will be able to provide the required airflow at as low fan speeds (read: low fan noise) as possible.

A side note on the increasingly fashionable silent cases in which the graphics card is installed vertically in front of a transparent side panel — most likely to be able to watch the (RGB) fans at work. This kind of the assembly I would advice against at present with most cases. The reason is that with this type of installation, the graphics cards are very often operated with very little distance to the case side panel and therefore the cooling of the graphics card suffers. And what does the graphics card have to do if one prevents it from access to the air required to breathe (cool): it will increases the GPU fan speed (read: the fan noise), in order to nevertheless still guarantee a sufficient air flow. So, from the point of view of noise, such a GPU installation is therefore not recommended, except for a few cases that really leave a lot of space between graphics card and side panel.

Although this specifiy possibility of noise reduction is certainly not an option for everyone and every use case, but it certainly makes sense in some situations and then it will least ensure an completely silent user experience at the location you are using your computer from -- even though the device itself might be as noisy as before. What I am talking about, you ask? I was referring to the option to reduce the noise of a computer by putting a sufficiently large distance between one's ears and the source of the noise.

One possibility is — assuming an appropriately designed apartment with the necessary space — to move the computer itself to a different room and to reroute the peripherals and the monitor cable to the place from which you want to use the device. Depending on the signal type, distances up to 5m (USB) or 10m (HDMI/DisplayPort) can be bridged without any problems, although this also strongly depends on the quality of the cables and of course — in the case of your monitor — also the display resolution.

So, if an relaxed gaming experience that you can you want to enjoy from the comfort of your living room sofa is your main purpose, then putting your computer in a different room is surely a good option. This can be an adjacent room where your only route the cables though the wall to keep their length within the allowed limits. But if necessary you may even move your gaming PC even further away and only stream image and peripheral connection to your TV over a network connection. This can be done, for example, with Steam Link, whereby the corresponding clients are available in most modern SmartTVs as software applications. If this is not the case, you can still buy the SteamLink box and use this one to for game streaming. There is one limitation that one has to consider and that is the maximum screen resolution. In the case of SteamLink you are limited to a resolution of a maximum of 1920x1080. If you want more than 1080p for your game streaming, then there is another called Moonlight Stream, which also allows resolutions of up to 4k and up to 120 fps.
All of those solutions in which an image needs to be transmitted in good quality and high resolution require a LAN connection with sufficient bandwidth between the devices. In theory it is also possible to operate these combinations via WiFI, but this is definitely not ideal; unless you want to stream your games in 640x480 or intend to use this setup for more or less static content like Solitaire or Minesweeper only.

From my personal experience I can report that with both a SteamLink and a Moonlight stream solution, most games like Assassins Creed, Skyrim, or The Witcher can be run perfectly smooth and without any noticeable lag from the living room couch. What certainly works less well are more competitive shooters like Battlefield or Call Of Duty, but when playing those you usually sit in front of the PC with headphones and Teamspeak/Discord active anyway. And at least in the heat of those battles you most likely won't notice the increased fan noise anyway.

Date published: 16.06.2021

Author: Udo Rietschel

Written by Udo Rietschel

Udo is one of the two founders of Argotronic GmbH. He is writing software since he got his hands on his first computer (a ZX Spectrum 48k back in 1988, while living in the GDR; a country that does not even exist anymore -- much like his ZX Spectrum).
Today, he is working as a software engineer, creating software for autonomous vehicles as well as Argus Monitor, which is — in his (biased) view — the best fan control software for Windows.