(Figure 1)
See the March report for the previous update. Our on-line reports have been erased or modified by an outside person in Beijing. Because of this, we have had to password protect this file. See Prof. Kroo if you need access.
The rotor design code is currently being modified to incorporate an improved viscous swirl correction and more accurate representation of section drag characteristics. The viscous swirl effects are based on two-dimensional INS-2d calculations. Average wake deficit velocity as a function of Reynolds number and distance from the trailing edge has been determined from the computational results for a single airfoil. The figure below depicts these results. This deficit velocity is applied to each spanwise station of the discretized rotor as an additional term in the tangential velocity component. Effects of viscous swirl are generally negligible for large rotors operating at high Reynolds numbers, but for the very low Reynolds numbers and high solidity ratios being considered here, viscous swirl effects become very significant. The addition of this correction should have a strong impact on both the optimal incidence angles and chord distributions.
Estimation of airfoil profile drag is improved by utilizing a database of two-dimensional section drag polars across a range of Reynolds numbers. This will provide a more precise estimate of section drag than the coarse fits currently employed. The current fits capture the overall trends in drag as a function of lift coefficient and Reynolds number, but cannot discern the detailed performance differences between various airfoils. With the new method, the effects of implementing different airfoils can be easily explored by providing a different database file for each section.
The rotors specially designed for 5-gram Mabuchi motors used in an electric airplane are commercially available. These rotors can generate about 20 grams at 20,000 rpm. Reversed rotors of the same type are also required in mesicopter, but not commercially available. Therefore, efforts were made to reproduce rotors that match the performance and to generate reversed rotors to meet our needs.
The coordinates of leading and trailing edges of the sample rotor were measured with the help of Zeiss Coordinate Measuring Machines by Precision Tool Distributors in Sunnyvale. Distribution of chords and incidences can be determined by the measured data. The NACA 4402 airfoil was first adopted and fabricated of polyurethane. This first trial showed NACA 4402 was too thin to retain the blade shape during operation. Therefore, the rotor was modified resembling the shape of cambered plates with thickness of 7.5% of chord. With further sanding at trailing edges, the reconstructed rotor can generate 20 grams as expected. The other issue observed during spinning was the stress concentration at the junctions. Stronger connections and rounded corners were appended to prevent fatal breakage. By mirroring the solid model in CAD, the reversed rotors can be manufactured without difficulty.
Weight reduction was shown to be effective by mixing Micro-Balloons with polyurethane in last month's report. More materials were explored during this month. A simplified air-frame design, which is similar to final design without 15-degree tilt, was built to evaluate feasibility.
Besides polyurethane with Micro-Ballons (PU w/ MB), rigid polyurethane foam (PU foam) and white polyurethane with Micro-Balloons (wPU w/ MB) were cast into a block to determine density. The densities are compared as follows:
| Material | Density (g/cm^3) | Comment |
| PU w/ MB | 0.64 | Short mixing and handling time before cured |
| PU foam | < 0.15 | Expansion is not uniform and not controllable |
| wPU w/ MB | 0.43~0.51 | Long handling time before cured |
White polyurethane became more attractive than polyurethane because of longer handling time and lower density before and after mixing with Micro-Balloons. The drawback is that white polyurethane takes 12 hours to be fully cured, while polyurethane only takes 4 hours. White PU with Micro-Balloons and PU foam were used for manufacturing a simplified airframe structure. The first test was based on wPU w/ MB for the main structure and PU foam for shrouds. The result is shown in Figure 1. The PU foam is not rigid enough to retain the required shape when thickness is small. The total weight of this version is 1.1 grams. Another experiment was performed to build the whole structure out of wPU w/ MB (Figure 2). The shrouds are stiff enough and the whole air-frame weighs 1.7 grams. Some cavities are observed in the air-frame due to some air bubbles trapped inside while casting.
(Figure 1)
(Figure 2)
My efforts in the last weeks concentrated on finding an alternative power supply for the smallest of the mesicopters ( 8 mm smoovy motor) we plan to build in next weeks. Supercapacitors (using the double layer technology) look as a promising alternative to the rechargeable batteries for a short demonstration of a self powered flight.
The rechargeable batteries usually weight much more than the double layer capacitors and also it takes much longer to recharge them than to recharge the capacitor. Also the capacity of a rechargeable battery decreases gradually with the increasing number of recharge cycles. After 500 to 1000 life cycles the rechargeable batteries lose most of their storage capacity. The double layer capacitor could be charged and discharged for more than 1 million times without any change in its energy storage capacity. An import advantage of the double layer capacitor is its high current charge and discharge capacity. After being fully charged the terminals of the capacitor could be even shorted without damaging the capacitor. On the other hand the capacitor shows a linear discharge characteristics compared to a flat one of a rechargeable battery and for any given size, the energy storage capacity of the rechargeable batteries are still many times greater than the double layer capacitors of the same size.
I have located several manufacturers of supercapacitors ( Panasonic, Tokin, Elna, Maxwell Technology) compared their products and ordered samples from the most promising ones. The supercapacitors usually have low voltage ratings ( 2.5 or 5 V). This requires to combine several smaller ones in a series in order to satisfy our power requirements ( 8-10V/600 mA). Currently tests are being performed with a smoovy motor to test how long could supercapacitor supply the energy for the mesicopter.
Here is an image of the controls testbed by Gary Fay. The device has been powered up and testing is underway.
We are designing a 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.
Currently Chad Partridge is programming the PIC processor to send a dedicated pulse width signal. The pulse widths will be based on information stored locally and updated via RS-232 communication via a local PC (eventually the Pentium based machine).
Last update: 15-May-00 7:24:19 PM
WebEdit servlet by I. Kroo