NoiseBoard FAQs (frequently-asked questions)

Well kind of... if you want a typical guide, then look at Chris Ruckman's guide. This version puts a different perspective on the subject.

Sure it works. *BUT* it only works in a few applications were it is easy to control the sound field (or vibration) reliably, such as in headphones, small diameter ducts like car exhausts, air-conditioning ducts, aircraft cabins to some degree. For some applications, such as satellites, it is vital to have active vibration control systems on solar panels because they are lightly damped and will continue to vibrate for 20 minutes if they don't have an AVC system installed. Some sporting goods have active vibration control systems installed, and make claims of reducing vibration by 30% or 50%.

Skip to the end of this FAQ to read about more products or look at the links listed here for other products.

It seems that the ANVC community can be divided into the believers and non-believers. As you will figure out, I fit somewhere in between. Yes - it can work, but before you entertain any fanciful ideas, will passive control work?

The other important aspect is the human perception of the results. It is possible to achieve large reductions in tonal noise, yet to the lay person, the noise level sounds the same because of the broad band background noise, or even worse because a listener can focus on another tone. I can cite two examples:
- A large helicopter manufacturer spent many millions of dollars developing an active noise cancellation system for their helicopters, and achieved an impressive 20dB reduction at the main gear mesh frequency. However in the wisdom of the execs, it was decided not to implement the system because it didn't sound much quieter. The noise inside the helicopter cabin also came from harmonics of the gear mesh and the main rotor noise, so a passenger didn't really think it was much quieter.
- Another example was from tests on an electric vehicle. Listener tests showed that by removing a high amplitude 2kHz tone unmasked another tone at 8kHz that was lower in amplitude but substantially more annoying.

The facts are that humans focus on the most annoying noise. If an annoying noise is elimated, attention is drawn to another annoying noise.

So even if the active control system works, a lot of work needs to be done on the perception (or psychoacoustics) of the noise to ensure that people perceive that the noise level is quieter *and* less annoying.

There are several mechanisms in which active control is used to reduce noise and vibration.

Cancellation
The easiest mechanism to understand is destructive and constructive interference. Imagine throwing a stone into a pond and you'll see that there are ripples in the water. If you threw another stone near the first stone, you'll see that there are two sets of ripples in the water. At some locations where the waves meet, they will interfere with each other and cancel so that the amplitude of the wave is zero. At other locations the waves will constructively interfere and the the amplitude can be twice as big at the original wave! As some of the waves are double in size, and some are cancelled, the total energy is conserved. This type of control is called "local cancellation".

Hence you now understand why active control in enclosures is very complicated. At some locations in the enclosure it can be very quiet, yet at other locations it can be twice as loud!

Impedance Control
There are two types of impedance control that can be used. The first type is where all the noise is absorbed by the control system. Imagine that there is a noise source (call it the primary source), such as a loudspeaker in a box that generates noise out of one side of the box, and another loudspeaker in a box is placed facing the front of the primary source, and is used to reduce the noise from the primary source, and we'll call this the secondary or control source. As the diaphragm on the primary source moves outwards, the diaphragm on the control source can be made to move inwards at the same time, effectively "sucking in" the sound energy. In order to get the diaphragm on the control source to move, it is necessary to apply power to it, to overcome friction and the inertia of the diaphragm. The effect should be that there is no sound radiated from the primary source. What has occured is that the secondary source has changed the ability of the primary source to make noise, which is called a change in the "radiation impedance".

Imagine holding a long piece of rope that passes through a wall, and the rope is rigidly fixed to the wall. If you quickly whip the end of the rope and send a wave travelling down the rope, you will see that when the wave reaches the wall, the wave will be reflected back to where it came from, and on the other side of the wall the rope will remain stationary. Now imagine that instead of a wall holding the rope, that a person was holding the rope and was able to provide the same reaction forces as the wall.

The next obvious question is, "but what if there is a reflection at the source end, won't the energy just keep sloshing back and forth?" The answer is yes it will, and the internal damping of the system will dissipate the energy.

Both these types of control can be used in the active control of noise travelling along air-conditioning ducts or exhaust stacks (big chimneys).


Active Structural Acoustic Control
A much more complicated mechanism involves "modal re-arrangement". A vibrating plate has many vibration "modes" that contribute to the overall response. By applying control forces to the plate, it is possible to re-arrange the energy that is contained within the modes so that the amplitude of vibration is less, yet the total vibration energy in the plate remains the same.

It has been found that the least noise radiated by a plate does not occur when the vibration of the plate is minimized at a number of points on the plate (except if the entire plate can be made completely still). The least noise is radiated when the plate is still slightly vibrating in such a way that the vibraton modes are poor radiators of sound, or that the modes combine to cancel one another.

A bit of a history lesson is in order.

Over the past decade there have been numerous companies that have been set up only to collapse. Companies such as Digisonix, Walker Exhausts, NCT and many others, have pumped millions of dollars into the development of products. Some companies have survived, and many have closed business. The problem is that it is difficult to make it work reliably and some applications are very complicated. Apart from the technical problems, there have been very intense legal battles.

In 1984, a company call Noise Cancellation Technologies (NCT) was formed and started to develop commercial products using active noise and vibration control techonology. There were ongoing legal battles with regard to patents and ownership, so in 1987 one of the key players, Tom Hesse left to start another company, Active Noise and Vibration Technologies (ANVT). The administrators were able to raise over $14 million in equity [see here]. The legal battles in between NCT and ANVT intensified and in 1994, the ANVT owner decided to sell his company to NCT. All the employees from ANVT were fired. There was an industry backlash from the fall out of these companies.

Perhaps the biggest winner of the development of active noise and vibration control are the patent lawyers. Patents have been taken out on active control algorithms that are the heart of any controller. The validity of these patents has yet to be tested however the lawyers are lurking around just waiting until the day comes where a company uses a patented algorithm or idea, so they can sue on the grounds of patent infringement.

For your own amusement, go to the United States Patent and Trademark Office web site and search the assignee name for all years for these companies "Noise Cancellation Technologies" 74 patents; "Active Noise and Vibration Technologies" 6 patents; "Nelson Industries", 25 patents with the word "active". The scope of some these patents are very broad (You can read a description of the patent at the USPTO website), for example

• 4,677,676, "Active attenuation system with on-line modeling of speaker, error path and feedback pack";
  
• 5,638,454 "Noise reduction system";
  
• 5,091,953 " Repetitive phenomena cancellation arrangement with multiple sensors and actuators"
  
• and many many more...
  

The intensity of litigations, and the broad scope of existing patents are formidable barriers to the advancement of active control technology. As one can appreciate, it is only worth suing another company if money can be gained from the battle, from either the protection of an income stream or to obtain a cut in the profits by licensing the technology.

Fortunately (?), it means that
universities are excluded as potential threats, because they don't make signficant amounts of money. Hence they can continue to conduct research, free from the legal battle grounds of the corporate world.

Despite the legal battles, there are commercial products that are sold with active noise and vibration control systems such as headsets, baseball bats, tennis racquets, skis, snowboards, car suspension, aircraft cabin noise reduction systems, deformable mirrors for telescopes, air-conditioners, stabilized camera lenses, and many others. See the links here for active control products. All these products use feedback type control. The patent hassles are with feedforward type control algorithms.

• Chris Ruckamans's FAQ is the most widely referenced formal discussions on the subject. Hopefully this one has provided a different perspective!
  
• There is a good article called, "Industry Overview of Active Control Methods and Applications", by Thomas L. Lagö that describes the active noise and vibration control industry, the players, the history, legal battles, and commerical products.
  
• Check the books section on this web site. If you want a good introduction to ANVC, then get either "Active Noise Control Primer" by Snyder or "Understanding Active Noise Cancellation" by Hansen (*). Both are introductory books, not intended for propellor heads. If you are after books with plenty of "meat" and Greek characters, see the other books on active control.
  

Before you jump down this path, are you absolutely sure that a passive control technique will not work? How about an adaptive-passive technique such as an automatically adjusting Helmholtz resonator or quarter-wave tube?

Fully active control systems are very complex and extremely difficult to make reliable for unexpected events.

I haven't talked you out of it? Okay, then you can buy controllers off the shelf that are relatively inexpensive that were made by Causal Systems. Look under the products section for Causal Systems and Arbor Scientific.

If you want to start from scratch, then I recommend getting one of the DSP evaluation borads from Analog Devices (EZ-KIT, EZ-KIT Lite) or Texas Instruments (DSP Starter Kit). You will have to write your own software, or adapt some from the internet. There is a great deal of free software that is available off the internet for digital signal processing. A book by Kuo and Morgan contains a disk with ANC algorithms written in C and assembler.