Scott Jensen and Dr. Scott Sommerfeldt, Physics and Astronomy

The problem of actively reducing noise from a small axial fan, a.k.a. a computer fan, is a complicated and difficult problem. A team of several BYU students have been working on the problem for about four years now. About a year ago, we started to solve the problem of reducing broadband noise from computer fans.

I started by researching as much as I could about active noise control: the reduction of noise using noise cancellation produced by electronic means. I read several articles from scientific journals on the topic of active noise control. Most articles addressed the topic of noise control in a duct, such as would be used in an air conditioning system. While this is related to what I was trying to accomplish with computer fans, the techniques are different because there isn’t any duct involved with computer fans. Active noise control of computer fans is a largely unexplored area of research.

With the help of graduate student Matt Green, we developed a feedback control method for reducing broadband noise from the fan. The main principle of a feedback system is that a signal is not only outputted from a system, but it is also fed back into the system. This feed back signal is used to check the output signal by making sure it is outputting what is desired.

A prototype system to test the feedback control was constructed. It consists of a rectangular aluminum metal plate, the size of the back of a standard desktop computer. A hole was cut in the back of the plate to accommodate the mounting of a computer fan. Four one inch diameter loudspeakers were also mounted in holes around the fan. These speakers were placed in optimal locations determined by a previous BYU graduate student, Kent Gee. Four small microphones were taped around the fan as well, so that they could easily be moved to the optimal positions.

The fan produces noise that is picked up by the microphones. The signals from the microphones are fed through digital filters, which filter out the higher frequency noise. The higher frequency noise is much harder to eliminate, so we filter it out, so as not to interfere with the band of noise we would like to reduce. I first designed analog filters for this purpose, but it was later discovered that the task could more easily be accomplished with available digital filters. The signals are then fed into computer controlled digital processors. The signal from the processors is then sent to the four speakers. If the system were ideal, the sound from the speakers would cancel out the sound from the fan and it would be completely quiet. The microphones would then not pick up any signal and the algorithm in the digital processors would not need to change. But in practice, sound is not always totally eliminated and so the microphones are the source of “feedback” to the digital processors. The microphone signals tell the processors how well it is doing to reduce noise. The processor will adapt its algorithm to make the signal from the microphone be zero or no noise. This system was tested and we were able to achieve a significant reduction in broadband noise by as much as six decibels. We hope in the future to achieve around twelve decibels of broadband noise reduction.

Throughout the course of the last year, BYU has gained more recognition for its work in fan noise reduction. As a result, several fan manufacturers are interested in sponsoring fan noise reduction research at BYU, which is great news for the program. Because of this increased interest, more prototypes needed to be built. Previously, the electronics used to reduce the tonal noise from fans was in the form of various scattered circuit boards. It was not very easy to quickly put together a working demonstration of fan noise reduction for a potential client. I volunteered to remedy this situation by designing one circuit board layout that would include all of these various circuits on one board. This board would be mounted in a metal box that would include all of the input and output ports and switches necessary to be compatible with different kinds of computer fans and setups. Although this part of the project did not deal specifically with broadband noise reduction, it did further the overall goal of noise reduction from computer fans.

I also designed a circuit board layout for a small microphone preamp circuit. This circuit is necessary to amplify the very tiny electrical signals generated by the microphones located around the computer fan. Both of these circuit board designs I made were mass produced by a printed circuit board manufacturer and successfully tested.

Excellent progress has been made over the past year that I have been involved with the fan noise project. I have now graduated and will be moving on to other things, but the fan noise project will continue at BYU. More improvements will be made in reducing broadband noise from computer fans, and the overall noise level a computer fans will continue to decrease through the use of active noise control methods developed at BYU.