Parts. It could be active system cooling with fans, hard drive - hard drive. It usually works very quietly, but if its sound annoys you, then you can change it to solid state SSD drives... This will greatly increase the speed of the subdisk system.

If everything is in order with your hard drive, it means that the cooling fans are the source of the noise.

Ways to reduce noise from PC coolers

Noise can be eliminated big amount ways:

Reduce the load on the processor;

Replace the cooling system;

Forcibly reduce the speed;

Replace small coolers with big ones;

Lubricate the fans;

Check the fan setting;

Clean the radiator from dust and dirt.

How to remove noise

The more the CPU is loaded, the more heat it generates and the stronger its fans work. And if the program also hangs on the computer, then the load will additionally increase. To fix the problem, it is enough to close the frozen program.

If you replace the cooling system with a liquid one, you can improve the heat dissipation. At the same time, during their operation, you will only hear the sound of a running pump and flowing water.

Exists special programs to reduce the rotational speed of coolers. One of the most popular is SpeedFan. It is good in that it can control the speed of the fans in automatic mode.

The larger the fan impeller, the better effect the cooling it provides. At the same time, the number of its rotations is less, and, accordingly, the computer noise is reduced. Fan bearings are also different. Among the two types - rolling and sliding, it is better to choose the first option, since it is more stable in emitting sound effects.

Sometimes it is enough to just lubricate the fan to make it quieter. For these purposes, sewing machine oil is suitable. If you want to lubricate the CPU cooler, pry off the sticker that is glued to the bearing. Put some oil in there and stick it in place.

The fan setting is very easy to check. You need to enter the BIOS of your computer and check if the Smart Fun Control options are enabled in the PC Health Status section. Depending on the type of motherboard, the name of the function may vary. It must be enabled, that is, have the Enable status.

If a lot of dust has accumulated on the radiator, this disrupts the operation of the system, the fans have to work with a vengeance. To clean them, there are special cans of compressed air, which can be purchased at a specialized store.

Knowing ways to reduce the noise of your computer, you can easily cope with this problem. But if your device makes a sound that resembles a squeak or whistle, then your PC needs a replacement part. Monitor the health of your computer or laptop and fix problems in time.

The closet or kitchen is designed to create optimal conditions for air exchange. A correctly installed and periodically switched on (manually or automatically) fan allows you to maintain acceptable level humidity, although it creates a certain noise. Sometimes the humming of the ventilation system leads to a feeling of discomfort and requires action to be taken to solve the problem. There are many reasons why the fan makes noise - and most of them can be eliminated on your own.

Noise and hum are signs of a problem

It is recommended that you pay attention to noise before purchasing a device. The value of the indicator is usually indicated in the instruction manual of the device. If the noise level is higher than the permissible values, it is worth giving preference to less noisy models, even if you have to overpay for improved performance.

The noise level of the device in the list of characteristics on the website

The noise level should also be checked after installing the fan. If the value of the noise load is significantly higher than that specified in the passport data, the problem may be an incorrect installation. You can avoid the problem by entrusting the installation of the fan to an experienced craftsman. Another reason for noise can be a factory defect - such a device should be returned under warranty.

The problem is also considered serious in cases where the fan hums, the noise from which was previously almost imperceptible. They solve it by removing the device and checking its condition. If the moving parts are heavily worn, the fan should be replaced.

Checking the condition of the impeller

The level of noise generated by ventilation should be compared with the current regulations. So, in the daytime in residential premises, the volume should not be higher than 40 dB, at night - 30 dB. Although, unlike the noise of other equipment (for example, a computer processor cooler), the fan in the bathroom, kitchen or toilet does not work constantly, which means that the maximum level can reach 70–80 dB - but not higher.

Causes of noise

To reduce the noise level from forced ventilation, you should choose models of popular brands that are distinguished by quiet operation and long-term operation. These include the Silent, Viessmann, Vortice and Maico brands. Even this equipment may not run as quietly as expected.

The main reasons for the increase in the noise level during the operation of the fans are as follows:

• increasing the level of vibration of the impeller;
• engine malfunction;
• increased bearing friction;
• too high air velocity through the impeller blades;
• violation of verticality or poor fixation of the device;
• poor assembly (one of the main problems when using cheap models);
• lack of preventive maintenance of the device.

Preventive work

To avoid problems with the installation and to do everything correctly, contacting a qualified specialist who can take into account all the nuances of installation will help. So, for example, only models with ball bearings are allowed to be installed on the ceiling, and too long the fan spigot leads to increased air resistance and loud noise.

The reasons increased volume the operation of the fan can become sound waves propagating through the duct. Working in normal mode, the device starts to hum a lot - first of all, this applies to channel models. The problem is no longer only in the fan, but in the whole system, therefore, comprehensive measures should be taken, including sound insulation of ventilation ducts.

Soundproofing methods for duct fan

The reasons for choosing duct fans include the need for forced ventilation over a relatively large area. Their characteristic feature is their installation not into one of the walls of the room, but inside the ventilation duct. Such devices are served by several rooms at once, and sometimes the whole apartment or house. higher, which also leads to an increase in the noise level.

Duct fan

One of the ways to solve the problem of noisy operation of the ventilation system is high-quality sound insulation of air ducts. This requires:

1. Prepare the necessary tools (silicone glue, soft rubber, long-handled roller).
2. Cut into strips of rubber suitable for the length.
3. Glue the inside of the channels with them, pressing it tightly against the walls.

The rubberized surface will increase sound absorption and eliminate most of the noise from duct fan... However, this method is not suitable for apartment buildings, the residents of which will have to solve the problem by soundproofing the wall closest to the duct and reducing the cross-section. Changing the dimensions of the channel accelerates the movement of air, which leads to spontaneous damping of the sound wave in the laminar air flow. The wall is insulated thin layer mineral wool or other porous materials.

Reducing the noise of wall-mounted models in the toilet or bathroom

A loudly operating fan in a household room does not cause such a feeling of discomfort as duct devices installed in the center of a house or apartment. However, it is still worth solving the problem if the noise from the ventilation system has noticeably increased during operation. In order to remedy the situation, they use the following methods:

• transferring the fan to work with a lower number of revolutions, which automatically reduces the noise from the passage of air through its blades;
• checking the correct installation of the device;
• installation or replacement of silencers;
• check the compliance of the device with operating conditions (and, if necessary, replace).

Fan replacement

A good option for creating optimal air exchange in rooms and maintaining a comfortable noise level can be called special silent models. The volume of their work does not exceed 25–26 dB. This not only does not interfere with the users of the ventilation system, but also complies with sanitary standards.

The principle of noise reduction in such devices is the use of special vibration isolators, which reduce the vibration of the rotating elements of the fan. The bearings of the device do not require maintenance and provide silent continuous operation for 20-30 thousand hours. With a permanently switched on forced ventilation system, the service life of the equipment reaches 3-4 years, with periodic use - more than 10 years.

Installing a silent model

The choice of a suitable device and special soundproofing methods will reduce the volume of the forced ventilation system. And, although it will not be possible to completely get rid of the noise, its level will correspond to the norms. To increase the efficiency of work on eliminating the noise load from a working fan, it is recommended to entrust them to specialists.

Noise Reduction Methods for Axial Fans

S. V. Karadzhi, N.E. Bauman Moscow State Technical University
Yu. G. Moskovko, LLC "INNOVENT"

Keywords: ventilation system, aerodynamic noise, harmonic, fan rotor

Noise is an important parameters of the majority of technical facilities, affecting their operating properties, environmental efficiency and competitive ability. The main noise sources in ventilation and air conditioning systems are fans. The article offers various methods for reduction of fan noise.

Description:

Noise is an important parameter of most technical objects that affects their operational properties, environmental friendliness and competitiveness... In ventilation and air conditioning systems, the main sources of noise are fans. The article proposes various ways to reduce fan noise.

Ways to reduce the noise of axial fans

S. V. Karadzhi, MSTU im. N.E. Bauman, [email protected] site

Y. G. Moskovko, LLC "INNOVENT"

Noise is an important parameter of most technical objects that affects their performance, environmental friendliness and competitiveness. In ventilation and air conditioning systems, the main sources of noise are fans. Often, the restrictions imposed on their noise levels are a decisive factor in determining specifications the object as a whole, therefore, much attention is paid to reducing the aerodynamic noise of the fans. What does science currently offer and what is implemented in the designs? The reader will find the answer to these questions in this article.

To minimize the noise of the ventilation system (without the use of sound-absorbing devices), several conditions must be met. First, the ventilation system must be designed in such a way as to have minimum aerodynamic losses. Secondly, it is necessary to select the type of fan (radial, axial) and then correctly select the fan itself for the design mode. Finally, recommendations for optimal fan layout in the system must be followed to ensure a uniform fan inlet / outlet speed profile. At the same time, it is desirable that the fan is low-noise. This article provides short review works on ways to reduce aerodynamic noise in axial fans and a number of designs of existing low-noise fans.

Aerodynamic noise can be caused by various types of sources (monopole, dipole, quadrupole). These sources have different origins, but they can be divided into two large groups: sources that cause broadband noise (in which all frequencies are equally represented), and sources that cause discrete (tonal) noise (radiation is concentrated only at some frequencies).

Sources of wideband axial fan noise include turbulent boundary layer noise on the blades; vortex noise associated with traces behind the shoulder blades. Discrete noise sources include rotation noise (load and displacement noise) associated with the rotation of the impeller blades; interaction noise associated with the interaction of the impeller with the fixed elements of the flow path. A large proportion of the fan noise can be noise associated with impeller imbalance, but since it is not aerodynamic, it is not considered in this article.

Reducing turbulent and vortex noise is a challenging task due to the fact that this type of noise is associated with the flow around the impeller blades. To reduce it, it is necessary to optimize the shape of the impeller blades in order to ensure uninterrupted flow around the entire length of the blade. However, in this way it is possible to achieve a reduction in noise to one degree or another, mainly at the design mode of operation of the fan.

In some cases, for example, if the blades have a suboptimal aerodynamic shape, the noise of the boundary layer can have discrete components. In this case, to reduce noise, blades with a serrated trailing edge are used (Fig. 1). It is interesting to note that there are also fans with blades that have sawtooth inlet edges, which are also presented as low noise in advertising materials.

One method of reducing broadband noise can be to design the fan at the lowest possible speed. It is known that turbulent noise is a source of the quadrupole type, and its sound power is proportional to ~ u 8, and the vortex noise is a dipole-type source, and its sound power is ~ u 6 where u- peripheral speed. With a decrease in the rotation frequency, the rotation noise also decreases, which has a dipole (load noise) and monopole (displacement noise) nature, and their sound power is proportional to ~ u 6 and ~ u 4 respectively.

There are techniques that make it possible to design axial fans at lower rotational speeds by increasing the aerodynamic load on the blades. For example, due to a number of measures, including a decrease in the design speed, the noise of one of the fans of the life support system of the ISS "Alpha" was reduced by 8 dBA. Since in this case the levels and distribution of pressure on the impeller blades and, accordingly, the broadband noise change, this method does not always lead to the expected result.

Discrete components of the acoustic spectrum associated with rotation and interaction noise, as a rule, have higher levels by 15–20 dB than broadband turbulent and vortex noise. Therefore, discrete noise has the most annoying effect on humans.

One of the directions for reducing the rotation noise, which is actively developing at present, is the use of impeller blades with a curved axis of alignment of the profiles. In fig. 2 shows a blade and signals of rotation noise from its various sections (having a different phase due to the spatial shape of the blade). On the right is a vector diagram of the sum of the signals, from which it can be seen that with the correct combination of phases and amplitudes of the signals, the rotation noise can be greatly reduced. The idea of ​​forming a phase shift acoustic waves from various sections of the impeller blades by changing the shape of the alignment axis of the profiles is presented in. As an example, Fig. 3 shows a fan wheel with a profile alignment axis bent in the direction of rotation.

This effect is widely used in axial fans of remote condensing units of split systems, in which the entrance edges of the blades have a pronounced beak-like shape at the periphery.

In fans made according to the "wheel (K) plus straightening device (CA)" scheme, due to the interaction of the blade rows with each other, the spatial shape of the CA blades has a great influence on the fan noise level. So, in the works and the results of studies on the influence of the inclination of the blades of the SA are presented, at a certain inclination of the blades, a decrease in noise is noted. It should be noted that at present there are conflicting data on the influence of the spatial shape of the alignment axis of the profiles on the noise of axial fans, therefore, the appearance of the curved blades does not always indicate that the fan is really low-noise, as stated in advertising materials.

A method is also used to reduce the rotation noise by installing the blades with an uneven pitch. With an uneven step, a sequence of sound pressure pulses will be emitted from each of the blades at uneven time intervals, which leads to a decrease and "blurring" of discrete components (Fig. 4). The effect decreases with an increase in the number of impeller blades. The most widely used wheels of this type are found in the automotive industry.

In fans with an inlet guide vane (VNA), there is an interaction noise that occurs when the traces or other elements of the flow path facing the impeller interact with the rotating impeller blades. It is worth noting that the opposite effect also takes place, that is, the impeller affects the VNA, similarly, the straightening apparatus (CA) affects the impeller, since such interactions propagate upstream.

The ratio between the number of blades in the VNA or SA and the wheel is of great importance. To minimize interference noise at certain harmonics at different speeds, the condition must be met in accordance with

(1)

where M u = u k / c 0 ;
M ca = c a / c 0 ;
u k is the peripheral speed of the ends of the blades;
c a - axial speed;
To m is the wave parameter;
m = │iz PK + kz AP │;
k- coefficient running through the values ​​of all integers;
i- harmonic number.

In the same work, nomograms are given for choosing a favorable ratio of the numbers of wheel blades and apparatus.

The book offers a simplified expression for choosing the ratio of the number of impeller blades and straightening apparatus:

(2)

where n- rotation frequency;
D- impeller diameter;
with- the speed of propagation of pressure impulses.

It should be noted that when choosing the ratio of the number of blades in the apparatus and the wheel, it is impossible to reduce the noise level at all the harmonics of the blade frequency.

Constructive elements of the fan: engine mountings, the electric motor itself (if it is installed in front of the wheel), etc. - also have a strong effect on noise generation. Just like BHA or CA, they create vortex trails and turbulent flow before or after the impeller, which can lead to an increase in noise levels. The noise level is influenced by the arrangement of structural elements relative to the wheel, the ratio of the number of attachment struts to the number of blades, the distance to the wheel blades, etc. Many computational and experimental studies have been carried out on the topic of rotor-stator-interaction, from which it follows that there is no complete clarity in this issue.

Conclusion

a) at the fan of circuit K:

• at the inlet in front of the wheel at a distance less than the chord of the blade there is a grid, electric motor mounts;
• the number of motor mounts is equal to or multiples of the number of blades; electric motor mountings are located from the wheel at a distance of less than 0.5 chord of the wheel blade;

b) for a fan with VNA or CA devices:

• the number of blades of the apparatus coincides with the number of blades of the wheel or is a multiple of them;
• the blades of the apparatus are located from the wheel at a distance of less than 0.5 chord of the wheel blade.

Literature

1. Aerodynamic noise in technology // ed. R. Hickling, 1977.332 p.
2. Jay Patel, Kingston, N. Y., United States Patent, 4,089,618, May 16, 1978.
3. Mitrofovich V.V. Determination of the limiting design parameters of axial fans with high static efficiency // Industrial aerodynamics. M.: Mechanical Engineering, 1991. Issue. 4 (36). S. 260-280.
4. Sustin S.A., Mitrofovich V.V., Isakovich S.A. Development of an experimental low-noise fan // Abstracts of the XIII All-Russian Scientific and Technical Conference "Gas Turbine and Combined Installations and Engines", MSTU im. N.E. Bauman, 2008
5. Munin A.G., Samokhin V.F., Shipov R.A. and other Aviation acoustics: at 2 o'clock. Part 1. Noise on the ground of subsonic passenger aircraft and helicopters. M.: Mashinostroenie, 1986.248 p.
6. Harvey H. Hubbard. Aeroacoustics of flight vehicles, Volume 1, Noise sources // NASA Reference publication 1258, vol. 1, WRDC Technical report 90–3052, 1991.592 p.
7. Belamri T., Kouidri S., Fedala D. and Rey R. Comparative study of the aeroacoustic behavior of two axial flow fans with different sweep angles // Paper FEDSM2005–77242, Proceedings of ASME FEDSM'05, 2005 ASME Fluid Engineering Summer Conference Houston, TX, USA, June 16-23, 2005.
8. Jifu Lu Xinli Wei, Yang Li. Research on aerodynamics and exit flow field of skewed fan-rotors // Power and Energy Engineering Conference (APPEEC), 2010 Asia-Pacific. Pp. 1-4.
9. Bamberberger Konrad, Carolus Thomas. Optimization of axial fans with highly swept blades with respect to losses and noise reduction // Fan 2012, Senlis (France), 18–20 April 2012, 12 p.
10. Horoshev G.A., Petrov Yu.I., Egorov N.F. Fighting fan noise. M.: Energoizdat, 1981.143 p.
11. Brusilovsky I.V. Aerodynamics and acoustics of axial fans // Proceedings of TsAGI im. prof. NOT. Zhukovsky. Issue 2650. M., 2004.275 p.
12. Lee, J. and Nam, K. Development of Low-Noise Cooling Fan Using Uneven Fan Blade Spacing, SAE Technical Paper 2008-01-0569, 2008.
13. Lu H. Z., Lixi Huanga, R. M.C. So and J. Wang. A computational study of the interaction noise from a small axial-flow fan // J. Acoust. Soc. Am., Vol. 122, No. 3, September 2007. Pp. 1404-1415.
14. Sawyer S., Nallasamy M., Hixon R., Dyson R. W., Koch L. D. Computational Aeroacoustic Prediction of Discrete-Frequency Noise Generated by a Rotor-Stator Interaction // 9th AIAA / CEAS Aeroacoustics Conference and Exhibit 2003.18 p.
15. Woodward, R. P., Elliott, D. M., Hughes, C. E., and Berton, J. J. Benefits of Swept and Leaned Stators for Fan Noise Reduction // AIAA-99-0479, 1999.12 p.

Over the past two years, we have seen a technological breakthrough in the production of heatsinks for processor coolers: extrusion heatsinks with a cantilever ratio of 18 and higher have become widespread, vacuum brazing technologies have become commonplace, bonded / fabricated fins and folded fins, previously considered almost exotic. but basic principle, on which the operation of the coolers is based, remains the same - air cooling based on forced convection. And just in the part of this notorious forced convection, nothing fundamentally new has appeared for a long time: manufacturers are following the well-worn path of increasing the geometric dimensions of the fans, the number of blades and the speed of rotation of the impeller. As a result, a cooler equipped with a powerful 60x60x25 mm fan with an impeller speed of more than 6000 RPM becomes the main source of noise in the computer, completely drowning out other very "loud" devices, be it fans in power supplies, case fans, hard drives etc. Undoubtedly, this state of affairs insistently requires us to conduct not only thorough temperature tests, but also an objective analysis of the noise characteristics of coolers.

In recent Thermaltake products, we have already briefly touched on this topic and presented the results of our measurements, without, however, going into methodological details. Now we will take a closer look at all the main points related to the acoustic properties of coolers and give an answer to three sacramental questions:

• How to measure?
• How to measure?
• How to get a reliable result?

Well, let's get started!

Initial prerequisites

And we will start, perhaps, by clarifying the causes of the noise ( unwanted sound) during operation of fans installed in computer systems (as part of processor coolers or separately in a computer case). There are only two main mechanisms for the generation of fan noise, and, accordingly, this noise is usually divided into two categories:

• aerodynamic noise
• mechanical noise

Aerodynamic noise... If the main cause of aerodynamic noise, let's say, is trivial (rotation of the fan impeller), then the physics of this phenomenon is rather complicated. Therefore, I will not go into particular details, but only note that the source of noise in this case is eddies in a turbulent boundary layer appearing on the surface of the impeller blades. The intensity of the noise here depends on angle of attack and the speed of rotation of the impeller (the greater the angle of attack and the higher the speed of rotation, the greater is the intensity of aerodynamic noise). The aerodynamic noise spectrum of fans is continuous (broadband noise) and, as a rule, has a maximum intensity at the frequency:

F max = K * (V b / d * cosα),

where K is the coefficient determined by the fan configuration; V b - linear speed of the blade (m / s); d is the maximum blade thickness; α is the angle of attack.

An additional source of aerodynamic noise is obstacles at the inlet and especially at the outlet of the fan. In particular, such an "obstacle" is the cooler's heatsink. The main reason for the noise in this case is the same vortices in the turbulent boundary layer, only now the boundary layer is already formed on the surface of the radiator fins. The intensity of the noise here depends on the air velocity and the configuration of the obstacles.

Mechanical noise... As the name suggests, fan bearings are the source of this noise. There is an opinion among users that mechanical noise occurs only due to wear or structural defects of bearings and should be practically absent in serviceable fans. In real life, things are different: of course, there are no ideal bearings! :)

If we take into consideration a standard sleeve bearing, then both on the surface of the shaft and on the inner surface of the bushing there are necessarily microscopic cracks, cavities, etc. Obviously, in this case, friction arises in the shaft-sleeve pair, and you cannot do without noise. A certain noise contribution is also made by the lock washers, which rotate (more precisely, turn) together with the shaft.

As for the structural defects of the bearing, they can seriously aggravate the situation and significantly increase the noise intensity. The most significant of them in the case of a sleeve bearing is the imbalance of the rotor (impeller), which usually leads to the so-called ellipticity of the bushing (in the cross section, the inner surface of the bushing has the shape of an ellipse instead of a circle). This defect is the reason for the appearance of pronounced tones in the low and mid-frequency region of the bearing noise spectrum. At the same time, the intensity of the noise increases, and in a subjective sense it becomes very annoying. Also, poor-quality lubrication (or its insufficiency) and a large gap between the shaft and the bushing also have a very negative effect on the acoustic properties of the fan on the sleeve bearing.

Turning now to rolling bearings, their very design predisposes to noise. After all, this is a whole complex of rubbing parts: an inner and outer ring (clips), rolling elements (balls), a separator. Moreover, rolling bearings, in contrast to sleeve bearings, are very susceptible to external mechanical influences (impacts, falls, etc.). And, as a result, they have a rich "bouquet" of defects, which usually leads to a higher noise intensity. Therefore, there is nothing surprising in the fact that fans on rolling bearings, even in normal (good) condition, are usually 2-3 dBA louder than their "twins" on sleeve bearings.

Now we will consider our first sacramental question and determine what measuring instrument can be used in our research practice.

His Majesty Sound Level Meter

International standards defining means and methods for measuring noise appeared relatively recently - in the late 60s. But they were the result of long-term painstaking work of many, many researchers, who laid down their heads (in a figurative sense, of course) to the glory of the triumph of science. But there was something to work on!

The main problem on the way to obtaining correct quantitative estimates has become, so to speak, the human factor, because noise (and sound in general) is a psychophysiological phenomenon rather than a purely physical one. Therefore, for a quantitative assessment of noise, it was necessary to take into account not only the physical properties of the phenomenon itself, but also its perception by humans and the effect on the body. Indeed, the human ear, in terms of electronics, is a nonlinear transducer of sound vibrations and plays the role of a complex bandpass filter (even a whole complex of filters): the loudness of low-frequency, mid-frequency and high-frequency tonal sounds with the same sound pressure level in subjective perception will be different (the tone of the middle frequency seems louder than low tones and high frequency). It is quite natural that the answer to the question of how to take into account psychophysicist noise in its quantitative estimates could only be obtained empirically.

In the early 1930s, a group of American scientists carried out the most important practical studies of the dependence of the subjective loudness of sound on its frequency. The result of these studies is a family of curves showing the difference in sound intensity levels for pure tones that appear equally loud. Later, these curves were called volume contours (the second name is the Fletcher-Manson curves).

Rice. 1. Contours of equal volume

On the basis of contours of equal loudness (more precisely, contours corresponding to levels of 40, 70 and 100 dB), it was proposed to introduce into research practice three methods of frequency correction of sound pressure levels to take into account the peculiarities of human sound perception and obtain a simple one-number characteristics instead of full frequency noise analysis (in octave or one-third octave frequency bands) or in addition to it. Now these three techniques are called frequency characteristics of correction (weighting) A, B and C .

Rice. 2. Frequency response of correcting circuits A, B and C

It should be noted that characteristic A has become the de facto standard, and the results of measurements of sound levels corrected for this characteristic appear in the overwhelming majority of regulatory and technical documents. As for the characteristics of B and C, the first has sunk into oblivion, while the second is still used in some industries (in particular, in the study of the noise of jet engines and military equipment).

So, the first requirement for our sound level meter has been determined: the presence of at least a correcting circuit A. Well, this will not be a problem, since such a "gadget" is present in almost all sound level meters (it is not difficult to implement it in the "hardware"). Further, will it be enough for us to restrict ourselves only to the sound level L A, corrected for characteristic A, and refuse to carry out frequency analysis of noise? In general, it is enough if we only want roughly confirm (or refute) the compliance of a particular cooler with the established hygienic standards (why we have the right in most cases to "replace" the noise of the entire system as a whole with the noise of a cooler alone, I will tell you a little later). But our goal is not only that. A more important task for us is an objective comparison of the noise characteristics of various coolers, and in this case, without performing a frequency analysis of the noise (in octave or one-third octave frequency bands), such a comparison cannot even be hinted at. Therefore, frequency analysis simply must be an integral part of our experiment.

Well, one more, the second requirement for a sound level meter becomes clear: for our purposes, it is necessary to have in it technical means frequency analysis of noise. And here big problems (mainly, financial ones) may arise:

1. The most flexible frequency analysis of noise is possible only with specialized spectrum analyzers, which are usually monstrously expensive (cost only software tools processing the results of the experiment can number more than one thousand "evergreen").
2. In practice, it is usually limited to analyzing noise in octave bands, and most modern precision sound level meters have built-in octave bandpass filters to enable this analysis. Sound level meters with built-in octave filters are, of course, cheaper than spectrum analyzers. But their price is in the range of 5-10 thousand, which, as you know, do not lie on the road.
3. In some cases, it may be necessary to analyze noise in one-third octave frequency bands. Not all sound level meters have filters that allow such an analysis and are often an option supplied by separate order. The most interesting thing is that this "option" usually costs the customer a pretty tidy sum and in very "neglected" cases can be at least 70-100% of the cost of the sound level meter itself!

And, finally, there is one more, already the third requirement for our measuring equipment: it must be accurate and have good parameter stability. Problems may also arise here, since not all (even relatively expensive) sound level meters are equipped with high-quality high-sensitivity microphones and have a really low level of intrinsic noise introduced by the measuring path.

Yes, there are a lot of problems. But they still needed to be solved somehow. I will say without undue modesty: we managed to do it, and without any significant losses both in quality and quantity ;-)

We did not pursue advanced measuring technology, but opted for the "old man" Bruel & Kjaer Type 2203, which is a reliable analog device that has successfully "plowed" nearly twenty years of experience without a single comment.

Why Bruel & Kjaer Type 2203 sound level meter? Because this device:

• fell into our hands on the most acceptable conditions ;-)
• corresponds to 1 class of accuracy according to GOST 17187-71 and is entered in the State Register of Measuring Instruments
• allows for on-line calibration with an internal reference voltage source
• the quality of the measuring path is not much inferior to the most modern sound level meters from Bruel & Kjaer and Larson Davis

There is another very important point, which played a decisive role in the choice of this device: our sound level meter was part of, so to speak, a VIP set. And it came to us in its composition, which includes, in addition to the sound level meter itself, additional sets of octave and one-third octave filters - Type 1613 and Type 1616, respectively.

As a result, with the involvement of a precision sound level meter Bruel & Kjaer Type 2203, all three of the above requirements for our measuring equipment were almost completely satisfied.

Of course, only one measuring instrument (even the most modern and high-precision one) will be a useless toy without a well-calibrated measurement procedure, in other words, without a well-thought-out and well-organized experiment. And, as you understand correctly, we are talking about the time to consider our method of measuring noise and answer the second sacramental question :)

Setting up an experiment

The procedure for correct noise measurements is significantly complicated by the fact that they require a strictly defined acoustic environment (measurement conditions), be it a method for determining the sound power level of noise sources in free sound field or, conversely, in diffuse sound field ... The only method that does not depend on external conditions when making measurements is to determine the sound power level based on sound intensity ... But its implementation requires a specialized sound level meter equipped with a two-microphone intensity probe. We simply do not have such a sound level meter at our disposal.

Therefore, based on the capabilities of our equipment (and our own capabilities, which do not always coincide with our desires :)), when choosing an experimental technique, we settled on method for determining the noise characteristics of noise sources in a free sound field above a sound-reflecting plane(GOST 12.1.026-80). Why was this particular method chosen? There are several reasons:

At first, this method not very picky about the measurement conditions. The experiment can be performed both in semi-muffled chambers and in open areas and indoors.

Secondly, the microphone of our sound level meter has an optimal (linear) frequency response precisely in a free sound field.

Third, this method allows us to limit ourselves to frequency analysis of noise in octave bands instead of analysis in one-third octave bands. For our purposes, in most cases, frequency analysis in one-third octave bands will be unjustified both in terms of the time spent on it and the quality factor of the result.

And finally, fourthly, we have access to a semi-muted camera.

Now, briefly about the measurement procedure itself (all the details of such measurements can be found in the text of GOST). The experiment is carried out in a semi-muffled chamber (a muffled chamber with a sound-reflecting floor) with geometric dimensions of 5x5x4 m.Before measuring the noise level of coolers, the background noise level is estimated (measured in the center and along the perimeter of the room at four points at a distance of 1 m from the walls, the results are averaged) ... Then the coolers are fixed in the center of the room at a height of 0.35 m on an elastic suspension mounted on a low tripod. A hemisphere with a radius of 1.2 m was selected as the measurement surface, and the number of measurement points and their location on the surface of the hemisphere meet the requirements of GOST. Initially, the sound level L A is measured at each point. Based on the averaged result, a decision is made on the possibility of further measurements or on the need to make corrections Δ to the measured sound levels (sound pressure) in accordance with the conditions of Table 1.

Table 1

If the difference ΔL is more than 6 dBA, then a series of measurements of the sound pressure levels in octave frequency bands and the sound level L A is carried out at each point; each measurement takes 3 minutes and the average value of the readings is recorded. The working results for all points undergo further mathematical processing (analyzed and averaged) to obtain the final research result - corrected and averaged sound pressure levels in octave frequency bands or L A sound levels. Sound power levels are not determined, but if necessary, this procedure can be easily carried out based on our final research results.

So, it seems, it's time to start considering the methodology for processing measurement results and answer the third sacramental question.

Initially, the array of measurement results is analyzed, and according to the conditions of Table 1, the necessary adjustments are made to take into account the background noise. Then the results are averaged according to the formula:

Where L m is the average sound pressure level in the octave band (or the sound level L A); L i - i-th level sound pressure in the octave band (or sound level L A); n is the number of measurement points; K is a constant that takes into account the effect of reflected sound (experimentally definite meaning of this constant is 0.9 dB, it is rounded off to 1 dB in the calculations).

Domestic GOST is limited to presenting the measurement result only in the form of L m. However, a related foreign standard (ISO 3744) insists on presenting the result in a slightly different form:

L d = L m + 1.645 * σ r,

where L d - protocol result (final result); σ r - standard deviation of measurement results.

The addition to the L m level actually takes into account the measurement error (I think the 1.645 factor is well known to metrology specialists). For our measurement method, the value of the parameter σ r, defined by the ISO 3744 standard, is 1.5 dB. We took some liberty and slightly increased the value this parameter(sometimes it is better to exaggerate a little than to underestimate the measurement error). As a result, the relationship that is used to represent the measurement result looks very simple:

L d = L m + 3.

The obtained L d values ​​are rounded to the nearest integer. The result of processing the results is a diagram, which is published in reviews.

Additional analysis

“Okay,” the most corrosive and critical reader might argue, “all this is good. But on what basis do you measure the noise of a cooler alone, separately from computer system in general, and after that compare the results obtained with the remote control, which are the hygienic standards of the general noise of the computer, and not of its individual components ?! "

I do not exclude that such critical sentiments might not have arisen among our readers; nevertheless, the question of the legitimacy of "replacing" the noise of the entire system with the noise of only a cooler is extremely important and requires consideration. Well, let's get this case straight!

Naturally, the end user would be interested in what the noise level in his particular system will be when installing a particular cooler. But it is not possible to give such information (moreover, objective and accurate). Let's dig a little in the prices of the offices retail accessories. And what will we see there? At least a thousand names of various motherboards, hard drives, video cards, ATX cases, finally! But all these components have the most direct effect on the overall noise level of the system, and when replacing, say, a hard disk or a case power supply, the level of this noise can change significantly. It is simply unrealistic to cover the entire spectrum of possible configurations - even Sisyphus would not have dared to carry out such measurements! ;-)

There is, of course, the methodological principle of the worst case: we first select the noisiest computer system and carry out measurements on its basis. The result obtained in this case will show the highest noise level of all possible and can be considered a completely objective starting point for further noise assessments of more "quieter" systems. But how to choose this notorious worst (in the acoustic sense) option from all the variety of configurations? There is no answer to this question, since the noise level of the system depends not only on the system itself, but also on the cooler installed in it. It is about structural vibration mentioned at the beginning of the article. The fact is that a cooler is not only a source of noise, but also a source of vibration. Vibration vibrations (which usually range from 10 to 500 Hz) are transmitted to the case through rigid joints (cooler fasteners, motherboard fasteners) and cause additional noise with frequencies up to 4 kHz and higher, depending on the case design (due to, so to speak, harmonic multiplication of oscillations). Therefore, it is likely that enough quiet system can seriously pump up in the acoustic sense when installing some other cooler with more high level vibration.

The situation is, of course, not easy. But a way out of it was found! We did not restrict rigidly to methodological principles, but carried out additional research, choosing several systems in four different cases (two branded and two cooperative-Chinese) and two relatively "vibroactive" coolers - GlobalWin FOP38 and Thermaltake Mini Copper Orb.

The results of the study turned out to be quite interesting:

1. The L A sound level of the system without a cooler (instead of it a Thermalright SK-6 copper heatsink was used) did not exceed 43-45 dBA (even in the Asustek FK600 case).
2. When installing the Thermaltake Mini Copper Orb cooler, the sound level of the entire system was 49-52 dBA (depending on the case), i.e. increased relative to the noise of the cooler in its pure form by only 1-4 dBA.
3. When installing the GlobalWin FOP38 cooler, the sound level was 54-56 dBA, i.e. decreased relative to the noise of the cooler by 1-3 dBA!

Based on the results of additional frequency noise analysis carried out for each case, we came to the following conclusions:

1. Although most users believe that computer cases are some kind of resonators that increase noise, this is really not the case in all cases: for coolers with excessively high noise levels (more than 55 dBA), noise attenuation is observed!
2. The housings tend to exhibit the properties of a band-pass filter (rather, a low-pass filter) - the sound pressure levels in one-third octave bands with geometric mean frequencies of 5000 Hz and above (and for "thick-walled" branded housings - and from 3150 Hz) turned out to be below the corresponding levels for a "just" cooler at least 1-2 dB.
3. The levels at the lowest frequencies, on the contrary, turned out to be "pulled up" by a maximum of 5-6 dB. This effect has largely manifested itself in branded cases.
4. At medium frequencies, the situation was ambiguous: Chinese cases raised the sound pressure levels by about 3-6 dB, while branded ones left them practically unchanged (increase within 1 dB) or even decreased.

So what do we end up with?

First, the sound level L A of computer systems filled with coolers with high-performance fans practically does not differ from the sound level L A of these coolers themselves (within the measurement error specified in the section Processing and analysis of measurement results)! Therefore, we have every right to compare our results with hygienic noise standards (however, this comparison is only indicative ).

Secondly, when the coolers are installed in the cases, the spectral composition of the noise changes: its concentration is observed in the low-frequency and mid-frequency regions.

Finally, thirdly, "thick-walled" branded enclosures are subjectively preferable to cooperative-Chinese ones: for systems in "left" enclosures, the noise is biased and amplified in the mid-frequency region of the acoustic spectrum, respectively, it seems more annoying than the predominantly low-frequency noise of the systems in branded enclosures, despite almost the same LA sound level in some cases.

Well, the answers to the three sacramental questions formulated at the beginning of the article have been given. You can make final conclusions with a more or less clear conscience ;-)

conclusions

Our method almost completely meets the requirements of GOST 12.1.026-80. Thanks to this, we obtain reliable and reproducible noise measurements, which allow us to carry out an objective comparative analysis of coolers in terms of their noise characteristics. Moreover, based on our results, we can give rough estimates of the noise and the entire computer system as a whole in the case of using coolers equipped with high-performance fans. As for constructive criticism of our methodology, it is, as always, only welcome! ;-)

In preparing this article, materials from the book "" / Ed. Barry Truax, Second Edition, Cambridge Street Publishing, 1999

Evenings all the best. Another article is on the way. Today the topic will concern the physical side of the computer, namely the noise of the fans. While working in technical support, I noticed that a lot of users ask to reduce the noise from their computer. First, find out the cause of this increased noise. It is possible that the computer has simply not been cleaned or oiled for a long time. But there is also a second option for this reason - they have flown / changed / reset the settings in the BIOS. Regarding the first part, it is not difficult to take a vacuum cleaner and a rag and remove all the dirt and dust. Regarding the second case, some BIOS navigation skills are required. Just now, I will tell you about this.

Reducing computer noise through BIOS.

So we made sure that the computer is clean and the fans are working. Now we turn on the computer and by pressing the special key, we get to the menu BIOS settings... You can find out which button to press by the start picture, or by the brute force method. The most common buttons are Del, F2 and F10.

In my case, the mother's ASUS board, in which the button to enter the BIOS - Del.

Immediately we find ourselves on the main tab -Main, for us there is nothing of value here, so with the left arrow we move to Power.

Having stopped on this tab, we begin to walk through the points and select "Hardware Monitor".

A page will open with a whole list of different add-ons. And it is also possible that the CPU Q-Fan Control item will contain the inscription Disabled, we can change this by going down to it and pressing Enter.

Well, and accordingly, it will now be revealed full list possible settings... Here you can find both settings for the processor fan and case fans, by the way case fans are designated by the word chassis.

Among all the presented, a very important point for us is CPU Fan Profile. The same profile with settings that defines maximum amount fan revolutions per minute.

Again, if you click on this item and a list (in my case) of three items will appear. I chose Optimal by default.

In short:

Optimal- this is an intermediate mode between productive and quiet;

Silent- this is the quietest mode;

Performance- this is a productive mode,

We opt for Silent. Then we press the F10 button and the BIOS will offer us to save the changes, we agree and the computer will reboot itself.

Already when switched on, the fans will first “whirr” strongly, and then decrease their speed to work in a quiet mode. If the noise still remains, then there are two reasons:

1. On motherboard there are two types of fan connectors. Some are signed as "CHA_FAN" and "CPU_FAN", the second is just "PWR_FAN". So only the first are controlled, the second type is simply feeding;
2. Conventional fans (with two or three pins) are connected, such fans are not controlled. The controlled ones have a connector of 4 legs.

Output.

After completing all the proposed instructions, we get a computer with a quiet mode of operation. Now you are calm and your colleagues do not complain. Good luck!