See the May report for the previous update.
The latest rotor design for the 5mm motor exhibits a significant improvement in performance over previous rotors. It is a 4-bladed rotor, as opposed to 5-blades in previous iterations, and the rotor diameter has been increased from 2.2 cm to 2.5 cm. The new rotor also implements the Re 6000 optimized airfoil instead of the NACA 4402 camberline used previously. This design utilized the new analysis code coupled with a non-linear optimizer. The figure below displays the required input power versus the experimental thrust for the new design and the best previous design. The 4-blade rotor obtains significantly more thrust for a given input power and reaches a higher maximum thrust. The displayed maximum in the figure is based on the power limitations of the motor and is not necessarily the aerodynamic maximum thrust.
The 4-blade design initially exhibited very low thrust values. This was likely due to aeroelastic deflections unloading the rotor blades. The smaller chords and thinner sections of the 4-blade rotor, combined with the use of an airfoil with higher pitching moment coefficients, caused significant structural deflections which were not seen on previous epoxy rotors. Early design experiments with polyurethane rotors did exhibit very poor performance, also due to structural deflections. The rotor hub diameter was increased to stiffen the structure and the attainable thrust improved tremendously. This simple design modification improved the performance of the rotor, but additional analysis is being done to accurately determine the nature of the problem to develop a more precise solution.
Test fixtures have been fabricated for both the 15 g and 60 g vehicles. The fixtures allow the vehicles to be operated with only two degrees of freedom, yaw and vertical translation. Pitch, roll, and lateral translation are constrained. This approach facilitates development of various control concepts, and greatly reduces the possibility of damage to either vehicle during control system development.
A prototype of the 5mm mesicopter was put together, see Fig. 1. It consisted of four 5mm smoovy motors, two right- and two left-hand propellers, four motor controllers and four 1 Farad super capacitors. First the four super capacitors were inserted into the frame. Second the capacitors were wired in series building one 0.25 Farad capacitor rated for 10 Volts. Third the controller circuit was soldered. Fourth motors were connected and wired with the controllers so that two motors rotate clockwise and two counter clockwise. Fifth the motors/controller blocks were glued to the frame. Sixth the controllers were wired to the super capacitor embedded in the frame and as last the propellers were glued to the motor shafts according to their rotation orientation.
The whole assembly process was very delicate and time consuming. The controller circuit consists of a single integrated circuit (IC) and three surface mount (SMD) capacitors. In order to reduce the weight of the controller the 3 SMD capacitors were glued on the chip and all the circuit connections were realized by soldering a thin isolated magnet wire to the corresponding IC or SMD capacitor pins.
Figure 1, Detail of assembled 5mm mesicopter.
The total weight of the mesicopter prototype is about 17 grams. This prototype could not take off with the power supply from the onboard super capacitors. With external power supply the mesicopter took of at 16 V input voltage. The current was about 1 A. The lift of a single propeller and of the whole mesicopter as a function of the input controller voltage is shown in Fig. 2.
Figure 2, Lift of the mesicopter as a function of controller voltage.
New version of rotor for 5mm Smoovy motors was designed by further developed analysis and design codes by Peter Kunz. The diameter of this new rotor is 2.5 cm. The cross-section is optimized (figure 1) instead of using NACA4402. Based on the designed geometry, the rotor was slightly modified for the purpose of manufacturability. These modifications included thickening the cross-section at the blade tip, strengthening the connections to the hub, and rounding sharp corners at roots. In order to avoid sharp geometry at leading and trailing edges, the building direction is reversed. The first trial is shown in figure 2. Due to smaller chords and larger diameter, the rotor performed large deformation during operation, resulting in lift fluctuation in the lift test. Therefore, hub diameter was increased from 2 mm to 5 mm to enhance stiffness in the second trial (figure 3). Test results of this version were included in the last month's report. Further modification of thickening inboard area (within 35% radius) was applied in the third trial, which was tested and used in the demo on June 27. Efforts were also made to manufacture both clockwise and counterclockwise rotors for Mesicopter use.
figure 1 cross-section of 4-blade design
figure 2 trial 1, small hub
figure 3 trial 2, big hub
Air-Frame Fabrication
The final air-frame design with 15-degree tilt (figure 4)requires using 5-axis CNC machine for manufacturing. The previous experience with building a simplified air-frame by 3-axis CNC machine helped us to identify material and manufacturing issues. From those experiments, white polyurethane with Micro-Balloon with effective weight reduction and sufficient stiffness was determined to be a suitable air-frame material. General fabrication sequences are expected to be similarly to the simplified version with further modification in 5-axis machine.
The airframe was fabricated in two sequential layers (figure 5, 6) with direct SDM (Shape Deposition Manufacturing). Part material is white polyurethane with Micro-Balloon, and support material is wax. Since the second layer is built on top of the first layer, the final part is in one piece without further assembly (figure 7). In order to increase the stiffness at the bottom of shrouds, a thin layer of polyurethane was added during the fabrication. The total machining time for the whole air-frame is about 6-8 hours, while the material curing time for each layer is 12 hours.
figure 4 air-frame final design
figure 5 first layer
figure 6 second layer
figure 7 fabricated part
A new prototype of the mesicopter has been completed. This prototype is lighter and smaller than the last. The longest dimesion is approximately six inches, and the total mass is 60.5 grams. The current motor/rotor combination will produce a total of 80 grams of thrust leaving extra thrust for climbing, maneuvering, and carrying payload. The counter- rotating rotors have also been manufactured and are now mounted on the mesicopter. An image of the latest mesicopter prototype is included below (Figures 3 and 4).
(Figure 3)
(Figure 4)
We currently have a simple feedback system that controls the vertical position of the 60 gram mesicopter on a shaft. We track a single LED on the mesicopter using an offboard vision system. This LED provides the position estimate. A simple PD-like controller provides the force command which is translated into a suitable pulse width command and sent to the mesicopter using a transmitter/receiver pair.
We are designing a more rigorous feedback system providing autonomous flight for the mesicopters. This system will consist of an off board binocular vision system, a pentium based computer for vision processing and control processing, a dedicated PIC microprocessor for transmission of the pulse width signal, and a transmitter/reciever pair for sending the pulse width velocity commands to the motors. The Pentium based machine will likely run the Linux operating system using RTAI (Real Time Application Interface) to assure hard real time response.
Peter Kunz and Ilan Kroo continued discussions with participants at the conference on "Fixed, Flapping, and Rotary Wing Vehicles at Low Reynolds Numbers". Interesting discussions with AeroVironment on their MAV work and with biologists Charles Ellington and Jeremy Rayner on flapping flight have suggested a few new directions that will be discussed in the final report.
Last update: 17-Jul-00 1:30:05 PM