Mechanical Design


The purpose of this project is to balance a ping pong ball with a produced air-stream similar to the demonstration shown with a hair-dryer. For the mechanical design, the objective is to create an apparatus that channels the airflow of a motor-driven fan with minimal loss in efficiency. The following is a documentation of the mechanical concepts and construction methods that lead to the finished apparatus. This project also serves as an introduction to the 3D printer and the various equipment and tools that are found in the SEng 466 lab in ECS 366.

CPU Fans: Early Trials

Initially, we wanted to use 12V – 0.25A CPU fans to balance the ping pong ball. This idea was quickly dismissed though, as the tests showed disappointing results; the fans were not able to provide the amount of airflow we required. We also experimented with stacking the fans and constructing funnels.  Unfortunately, none of these changes gave us a marked improvement.

Bernoulli’s Equations

In an attempt to predict the airflow before actual experimentation, we used Bernoulli’s Equation. From the equation, the pressure required to keep a ping pong ball in the air is simply:

 p= \frac{F_{p}}{A_{s}} = \frac{mg}{0.5 \pi D^{2}}

Funnel for the CPU fan designed with the 3D printer

Using the characteristic curves of one of the CPU Fan Specifications, the output velocity of a given funnel was calculated to be around 2.1 m/s. Using simple kinematics, the maximum height that can be achieved with this velocity was found to be approximately 20 cm, with an equilibrium height of about 10 to 15 cm. In the experiment, the results were drastically lower than the expected values. This is because in the actual apparatus, a funnel was built to concentrate the airflow to a smaller area. When that happens, the back pressure inside of the funnel increases. The performance of an axial fan against back pressure is very poor, and with the addition of turbulent air flow causes the overall efficiency of the system to drop significantly.

Stacking Fans in series

Experimentation with fans connected in series

Since one CPU fan could not provide the airflow required, one idea we had was to use multiple fans to force a larger airflow. According to the fan system source-book, stacking the fans in series doubles the pressure; as a result, it doubles the airflow output. However actual experimentation showed different results. The performance of the system fell short due to the angular component of airflow which is introduced by the rear of the fans. When the fans are attached in series, some areas of the system received better performance than the others, but overall the system only provided 10~20% better airflow and these values are not reliable.

Suggestions and recommendations

The thought of using CPU fans to balance a ping pong ball in mid air is a novel  idea but it is not very practical. The airflow provided by these CPU fans are just not sufficient for the project. Testing the CPU fans with Bernoulli’s equations showed inaccurate results. This is due to the turbulent airflow caused by the fan blades slicing chunks of air to produce airflow rather than using smooth a pressure differential. However, later in the project, new techniques were developed to improve the flow of air. With a more powerful CPU motor, and incorporating these new techniques, a CPU fan ping pong ball blower is possible.

Apparatus Design

motor and fan

Motor with fan blade assembly

After ruling out the CPU fan as an option, we searched for an alternative. The new motor that was chosen for the project is the 1811-2000 micro brushless outrunner which is the exact same motor that was used in the previous year’s quadcopter project. Since Bernoulli’s equation was shown to be inaccurate and something that could not be relied upon in practice, we chose to base our assumptions on the brushless motors on the fact that was once able to lift a quadcopter device, so it could surely provide enough thrust for our needs. One of the problems we encountered with the motor was that it was spinning in the wrong direction. In the quadcopter website, the previous group had the motor running in different directions to control the movement of the craft. The solution was to switch two of the 3-phase wires so it would operate in the correct direction.

Motor Mount and Base

The mount and base for the motor is a fairly simple design; it uses two pieces of Styrofoam as the base and a machined piece of aluminum to mount the motor and fan in place. The top piece of Styrofoam (coloured pink) has a slightly larger diameter hole than the bottom piece. This creates a ridge that allows the funnel to be later mounted on. Holes were also cut on the top piece of Styrofoam coincident to the aluminum plate mount. This is to allow the aluminum plate to rest flush on the surface of the bottom piece and also to provide an exit for the motor wires. The motor is mounted in this direction because a clearance of 10 to 15% of the diameter of the blades is recommended (see this paper on axial fan flow design). To allow better inlet airflow, the bottom edge of the base is recommended to have a bell inlet shape. Having a smooth curvature for the air inlet can improve efficiency by up to 10 to 15%. We sculpted this shape on the Styrofoam with a low grit sandpaper.

Cutting the styrofoam base with a hot wire

Base Mount for the motor and the assembly


The funnel was made using cheap poster paper that was purchased at a dollar store. We chose construction paper as it’s flexibility allows us to roll it into any kind of shape. According to the paper on axial fan flow design, the angle of incline of the funnel should be less than 15% in order to reduce the loss in airflow due to back pressure. It is also important to note that a larger reduction in diameter equates to a bigger loss regardless.

Homemade compass to draw large arcs.

The funnel was designed and modeled beforehand using the Solidworks CAD program. The “sheet metal” tool was used to flatten the 3D cone shape so that it can be printed on paper. The radius we needed to draw is quite large, so transferring the drawings on the the large sheet of construction paper required a very large compass. For this, we constructed a home-made one using two pencils and a piece of string. The string is measured to the correct radius before drawing the arc.

Left: The completed Funnel shape. Right: The fins that are fitted inside.

After the funnel was finished it was mounted to the base and tested. However, it still did not output the desired amount of airflow. The problem was that the air inside the cone was being recirculated. This caused turbulent flow which reduced the speed of the outlet air. We minimized this problem by inserting fins inside the tube, forcing the air to flow straighter and become more laminar.  Adding a long straight tube at the end of the funnel was also found to induce a more laminar flow of air. Adding the fins to the base also showed signs of improvement.

Sonar and Ruler mount

Ruler mount and sonar mount assembly

The mount for the sonar sensor was designed in Solidworks, which was later imported to a .STL file to be read by the Makerbot 3D printer. The mount was made in two pieces: one piece was the sonar mount, and the other was the base which houses the pins used as a hinge. There is a small screw drilled on the bottom of the base plate which allows the angle of the sonar to be adjusted. Electrical tape is wrapped around the tube of the funnel to provide a frictional grip that further secures the mount in place. The ruler mount was also made with the 3D printer after modelling it using Solidworks. It mounts on the rim of the funnel and the half circular extrusion on the edge secures the ruler in place. The ruler also acts as a guide for the airflow; the vacuum caused by the airflow pulls the ball to a higher height; we discovered this accidentally when we first mounted the ruler to the device.


Many times during testing, random objects around the table were accidentally sucked into the apparatus by the inflow from the fan. To prevent this, a safety net was installed on the base to prevent potentially harmful particles from clogging the fan blade.

Mesh guard to prevent harmful particles from entering in.

Complete Apparatus Assembly

The final assembly of the apparatus is very close to the proposed design. Several performance tweaks, such as the internal fins, were not expected but overall the apparatus designed in Solidworks fit closely to the final product.

Complete assembly of the entire apparatus

Final Thoughts

Overall the mechanical design objectives were achieved; the apparatus is stable and can channel airflow nicely with minimum loss in efficiency. The one drawback of this design is that the motor that was chosen produced too much noise at low speeds. The noise is so irritating that you want to keep the motor running at high speeds all the time. This noise can be reduced in future projects by choosing a lower noise motor, or by running the current motor at full speed and adjusting airflow by moving a shutter with a servo motor.

2 thoughts on “Mechanical Design

  1. sir, i am final year student and i want to make some different. please give me some ideas. thanks.

Comments are closed.