Search

Due to ever increasing consumer demand, today's marketplace is full of portable computing devices such as notebook computers, gaming machines, personal digital assistants, and smart phones. In each of these technologies, processing power and functionality levels are continuing to increase, resulting in increased thermal management issues. The former of these technologies already use an active cooling solution with a profile of 10-15mm while the latter two are quickly approaching the limits of passive cooling. Since a common characteristic in each of these technologies is the desire to reduce profile height, there exists a need to develop low profile cooling solutions that are capable of supporting this growing market. However, a prerequisite to developing such solutions is an accurate understanding of how the noise level characteristics of the driving fans scale with height. This paper experimentally addresses the acoustic scaling of low profile fans. The rotor heights range from 0.5mm to 15mm with diameters from 15mm to 32mm, giving an aspect ratio of height relative to diameter of 0.03-0.6. The results can be used as a foundation for the development of active cooling in mobile devices where both correlations for noise levels and basic design criteria are identified. [ABSTRACT FROM AUTHOR]

A piezoelectric fan is a flexible cantilever beam whose vibration is actuated by means of a piezoelectric material. Such fans have been employed for the enhancement of heat transfer by increasing the fluid circulation in regions which are otherwise stagnant. The main focus of past studies has been to predict and describe the heat transfer achievable using these devices, as well as the flow field generated by vibrating cantilevers. In order to directly compare these fans with their traditional counterparts such as small axial fans, the present work casts the performance of piezofans in terms of a characteristic often used to represent conventional fans, namely the fan curve. The primary focus of this paper is to determine the relationship between the pressure and the flow rate generated by miniature piezoelectric fans. Experimental measurements are obtained for fans with operating frequencies of 60 and 113 Hz. The maximum flow rate conditions yield nearly 30 1/min, while the greatest static pressure generated is found to be 6 Pa. The performance is highly dependent on both the vibration amplitude and frequency. Predictive relationships are developed to describe the experimental trends and provide insight into the sensitivity of pressure and flow rate to these operating parameters. These fans are directly compared to two commercially available axial fans, both in terms of overall performance and efficiency with which energy is imparted to the fluid. Piezoelectric fans are found to compare quite favorably using either of these performance metrics with a nearly order-of-magnitude increase in fan efficiency. A secondary focus of this paper is to explore the effects of fan installation details on fan performance. The proximity of surrounding walls is considered through the use of three different enclosures within which the fan is mounted. Effective inlet areas from which the air enters the fan are also identified. This paper provides a practical framework for determining the optimal placement and configuration for these fans in prototypical applications. [ABSTRACT FROM AUTHOR]

Piezoelectric fans are gaining in popularity as low-power-consumption and low-noise devices for the removal of heat in confined spaces. The performance of piezoelectric fans has been studied by several authors, although primarily at the fundamental resonance mode. In this article the performance of piezoelectric fans operating at the higher resonance modes is studied in detail. Experiments are performed on a number of commercially available piezoelectric fans of varying length. Both finite element modeling and experimental impedance measurements are used to demonstrate that the electromechanical energy conversion (electromechanical coupling factors) in certain modes can be greater than in the first bending mode; however, losses in the piezoceramic are also shown to be higher at those modes. The overall power consumption of the fans is also found to increase with increasing mode number. Detailed flow visualizations are also performed to understand both the transient and steady-state fluid motion around these fans. The results indicate that certain advantages of piezoelectric fan operation at higher resonance modes are offset by increased power consumption and decreased fluid flow. [ABSTRACT FROM AUTHOR]

The noise produced by cooling air passing through electronics packages arises from two sources. One source is the noise of the air-moving fan of either an axial or centrifugal type. This noise may have both tonal and random components and is strongly dependent on the way that the fan is placed in the unit and on where its operation is on the fan's operating curve. Often, fans are the dominant noise sources. The flow of air produces random noise due to the turbulence generated throughout the unit. Because the turbulent airflow is also responsible for heat transfer between the components and the air stream, we can regard this part of the noise as the irreducible noise due to cooling. If fan noise were eliminated, this part of the noise would remain. There is a relation, therefore, between the irreducible noise and the cooling of the unit. But the fan noise must also be considered. The relation between total airflow related noise and cooling requirements is developed in this paper for the irreducible noise. [ABSTRACT FROM AUTHOR]

In this paper, loss coefficients are determined for a fine and a coarse EMC screen placed in a swirling flow. In an axisymmetric swirl generator test rig the screen performance is measured under well-defined swirling conditions. Based on velocity and pressure measurements on one hand and on axial and tangential momentum balances on the other, directional loss coefficients are obtained. It is shown that in this way in-plane coefficients can be accurately determined. In contrast, the normal component can more accurately be determined in a basic wind-tunnel setup. Subsequently, the obtained coefficients are used in a commercial computational fluid dynamics code using a volume resistance model for the screen';. Validation of the numerical results against fan measurements shows that using in-plane coefficients is superior to one-dimensional models. Indeed, in the former case the effect of the screen on the tangential velocity component is well predicted by the model, whereas for the latter only the axial velocity components are affected. Finally, a full rack simulation reveals that component temperatures strongly depend on local flow phenomena. Therefore, it can be concluded that fan. induced swirl should always be modeled and that screen models should account for in-plane loss coefficients. [ABSTRACT FROM AUTHOR]