Businesses are developing their own standards and protocols for regulating noise in the workplace and defining the permitted noise levels generated by specific items of equipment.

The employer is required to take certain actions, such as providing hearing protection, information, training, and so on, when the daily exposure level is likely to exceed 80 dB(A). A quiet office exhibits about 40 – 45 dB(A) while a noisier environment of 60 – 65 dB(A) requires raising the voice while talking. The culprit behind this office noise is mainly the low humming generated by the cooling fans for personal computers and servers. The fans circulating air from an office HVAC system also contribute noise.

Before attempting to control noise, it is necessary to understand it. Sound waves travel through any type of media, including air, water, wood, masonry, or metal. Depending on the medium through which it travels, noise is considered either airborne or structure-borne. While airborne noise radiates from a source directly into the air and travels through it, structure-borne noise travels through solid materials, usually in direct mechanical contact with the sound source or by an impact upon that material. All structure-borne noise must eventually become airborne noise in order for people to hear it, but to achieve effective noise control, both airborne and structure-borne noise must be addressed.

Enlarge this picture Fig. 2. In a typical active noise control set up, a reference microphone captures the source noise and transfers it to a system controller, a digital signal processor that generates counter-phase signal and sends it to a speaker that transmits the anti-noise signal to the environment. An error microphone located at the point where noise reduction is required provides information to the controller, allowing it to make corrections and make the reduction signal more accurate. The reference microphone and speaker are typically located in a duct that is used to shape and simplify the unwanted noise source.

The basic idea behind noise control, sometimes called noise reduction or cancellation, is preventing the sound waves from getting to the ears. The traditional method of controlling fan-generated noise is the use of acoustic (sound absorbing) materials, known as passive noise control. It is common to find equipment lined with foam and where the cover can serve as a barrier. The foam is used to absorb acoustic energy, while the barrier prevents the residual noise from escaping.

There are two main limitations on using acoustic materials. First, these materials are less efficient on low-frequency noise, which can be the most irritating. The lower the frequency, the thicker and heavier the acoustic materials required. Second, optimal results require sealing of the noise source along ducts to allow absorption of the acoustic energy. Since thicker acoustic material inhibits ventilation, this form of noise reduction becomes unacceptable for electronic equipment where heat builds up. Longer ducts and sealing of the noise source creates unwanted resistance for air flow, and restricted air flow adds to the problem of heat dissipation, a critical issue in today’s increasingly powerful electronic equipment.

The limitations of passive control with acoustic materials can be overcome, however, by employing an alternative method that uses active noise control. With this alternative, one can generate a virtual barrier for noise without interfering with air flow into and out of the unit.           

Enlarge this picture Fig. 3. Silentium’s active noise control solution includes a virtual microphone that eliminates the need for an actual error microphone.

The basic principle behind active noise control has been known for more than 70 years. It involves the production of an opposing sound wave (anti-noise) that will cancel out the unwanted sound. This requires emitting a counter signal that is of the same amplitude, but 180 Deg out of phase from the signal to be cancelled. (See Fig. 1.) While conceptually simple, it can be complex to implement, and the level of attenuation is highly dependent on the accuracy of the system in producing the reductive signal (anti-noise) at the proper amplitude and phase.

In a typical active noise control set up, a reference microphone captures the source noise and transfers it to a system controller, a digital signal processor that generates counter-phase signal and sends it to a speaker that transmits the anti-noise signal to the environment. (See Fig. 2.) An error microphone located at the point where noise reduction is required provides information to the controller, allowing it to make corrections and make the reduction signal more accurate. The reference microphone and speaker are typically located in a duct that is used to shape and simplify the unwanted noise source. Reducing the complexity of the acoustic field by use of a duct makes it easier for the system to generate the proper noise cancellation.

Applying the theory of active noise control to the design of actual products poses some challenges. From a practical standpoint, there is a limit on size and weight one can put into a system to reduce noise. In addition, electronic equipment poses design constraints related to air flow due to thermal dissipation requirements. Furthermore, in the real world, one does not typically find the simple, ideal acoustic waves that lend themselves to effective cancellation by mirror waves.

Fig. 4. Illustration of a hybrid active/passive noise control solution for an axial fan.

There are several challenges to implementing practical active noise control in a point-to-zone application, where noise originates from a localized source, such as a piece of electronic equipment. One of those is timing, making sure that the unwanted noise signal and its opposite anti-noise signal meet at the same place at the same time. This objective is complicated by a number of factors, including a slight group delay in the dynamic speakers typically used for this purpose. One means for dealing with the timing issue is the use of a prediction filter, a set of mathematical algorithms that predict what the unwanted noise will be at the time the anti-noise signal meets it.

Another challenge is echo cancellation. This problem stems from the fact that the speaker generating the anti-noise signal emits it both forward and backward, so the anti-noise signal is picked up by the reference microphone and gets combined with the target unwanted noise signal. This issue can be solved by the creation of an echo cancellation filter within the controller, allowing it to focus only on the unwanted noise signal.

Another issue is whether to use feedback control, or feedforward control. With feedback control, the controller attempts to attenuate noise without the use of a reference microphone generating a reference signal. It uses only the error microphone. The problem with this approach is a limited frequency bandwidth of acoustic signals that can be controlled because of speaker dynamics. Feedforward control is generally more effective across a broader bandwidth, but requires the use of both a reference microphone and error microphone.

Silentium provides design engineers with a design and performance evaluation tool for noise and vibration reduction. Called the S-Cube development kit, it consists of an active noise reduction controller board, a calibration unit, two microphones and speaker. The controller allows the designer to create broadband noise and vibration reduction solutions, achieving up to 10 dBA reduction on top of what is achieved with acoustic materials, covering the full audible range. One of the microphones serves as the reference that captures and feeds the noise source to the controller, and the second is an error microphone that measures the residual noise during calibration in order to minimize it. The error microphone is not needed after the calibration process is done.

Silentium has overcome these obstacles to practical active noise control by developing a unique scheme, based on real-time, adaptive algorithms running on a digital signal processor. The proprietary algorithms cover a broadband audible spectrum ranging from 150 Hz to 1,800 Hz. A key innovation lies in the creation of a virtual microphone to replace the physical error microphone in a feedforward control system. This is accomplished by making a series of measurements with an actual error microphone in place, then applying those measurements to calculate an estimation of error signal that the actual error microphone would produce. Once the estimates are obtained, the actual error microphone is no longer necessary, as it exists virtually in the controller.

Eliminating the physical error microphone is highly beneficial, as it makes the active noise control solution product dependent instead of environment dependent. The product equipped with this solution can be placed in any location without the need to install error microphones. It can be easily relocated without affecting noise reduction performance.

The potential for this technology to reduce noise from electronic equipment has already been demonstrated. Silentium has introduced solutions that include an IT equipment rack-mount enclosure that promotes ventilation and noise reduction at the same time. Measurements by the manufacturer have shown this enclosure to reduce internal noise by the IT equipment and the ventilation system by up to 20 dB(A). Also, an add-on unit that was designed for a new Intel modular server has been shown to reduce the humming noise by 10 dB(A).

Residential appliances and HVAC equipment are also targets for Silentium active noise control technology. Typical appliance applications in which active noise control would be particularly pertinent include fans and air ducts/pipes, usually found in cooker hoods, air-conditioners and other ventilation appliances. By adding a duct, one can either improve the functionality of the device (as in the case of an axial fan), or not affect it at all (as in the case of air pipes).

For example, Fig. 4 depicts a hybrid active/passive noise control solution for an axial fan. A 90 mm x 90 mm x 20 mm axial fan was positioned at the midpoint of a 440 mm long duct with a square cross-section that was designed to avoid obstructing the airflow. The duct was lined with acoustic foam to reduce noise in the high frequencies. Two different active noise control systems were used, one on each side of the duct in order to control the noise emitted from the intake and output.

The hybrid example illustrates another point about active noise control, namely, that optimal noise reduction results can be achieved by employing both passive and active noise control methods. The passive acoustic absorbing materials handle the high frequency bands of noise while the active system addresses the lower frequencies.

For more information, email: info@silentium.com