Mesicopter Progress Report
September 1999


Summary

Work in the first month of the Phase II contract began on Sept. 1, 1999, although we have not yet been informed by the University that the contract is in place. Profs. Prinz and Kroo are loaning money to the project from other accounts for the time being so that we can get started with the aggressive schedule as proposed. Our basic goal is to develop two prototypes in parallel: a stability and control testbed that is much larger than the mesicopters, but uses the same configuration and control concepts, and a much smaller mesicopter, but still larger than our eventual goal. The large testbed will weigh about 60g while the smaller one will have dimensions of about 4 cm and weigh about 15g. These will allow us to develop the component technologies to fly our smaller (3 g) devices.

In each of the monthly progress reports we will provide brief summaries of work in the critical technology areas. Work will focus on certain aspects of development during particular months but we will include a summary of each area in every report. The general areas include: rotor aerodynamics and fabrication, motor system, batteries, structures, stability and control, system and application issues, and sensors and communications.

Rotor Development

Aerodynamics (Kunz, Kroo)

The initial rotor design for the 3mm Smoovy, when mounted on the motor, achieves the design rotor speed of 50,000 RPM only by drawing more current than anticipated. This indicates either a deficiency in the torque output of the motor or an increase in the aerodynamic drag from what was predicted. SEM and optical photomicrographs of rotors for the 3mm Smoovy motor appear to show discrepancies between the designed airfoil section geometry and the as-built part.

Airfoil sections that approximate the geometry from the photomicrographs have been evaluated using the INS2D incompressible Navier-Stokes solver. The initial results indicate that the as-built sections may yield a significant drag increase over the design airfoils. The as-built sections exhibit blunt leading and trailing edges and a constant thickness. From a manufacturing standpoint, the very thin trailing and leading edge regions of the design airfoils are problematic for the smallest rotors. Alternative airfoil sections are being explored which provide sufficient thickness for manufacturing, but may regain a significant amount of the lost performance.

Fabrication (Cheng)

Shelly will summarize this work. After the close-loop controller for 3mm Smoovy's motor was received, more lift tests were conducted. With this controller, the rotor could spin at targeted 50,000 rpm (9.0Volt and 92mA) and generate about 780mg lift. Different substrate(wax)-part(rotor) combinations were experimented. In phase I research, blue wax was used as substrate because of its nice machinability. However, only epoxy part had nice bonding to the blue wax during the machining. The other possible substrate candidates, yellow wax and purple wax, are tried with the part material of yellow polyurethane, less brittle and faster curing than epoxy. Both waxes are similar in properties, and can provide sufficient bonding strength during the machining. Parts are easy to be pulled out of the substrate without melting the wax. This would reduce the possibility to produce warpage or distortion due to thermal effect in melting wax. Some efforts are made to characterize the cross-section of rotors. Epoxy rotors, built with blue wax substrate, were embedded in yellow polyurethane; on the other hand, yellow polyurethane rotors, built with purple wax substrate, were embedded in epoxy. The embedded blocks were machined down gradually till the cross-section at the tip was revealed. The preliminary results from microscope showed the leading and trailing edges were not in the exact shape as designed due to the machining strategy adopted to remove flash at the edges. With the current method, the position of the cross-section in radial coordinate is not clearly defined; therefore, it's hard to identify if the chords are manufactured as required. Other methods to obtain more precise data will be tried.

Motor System

Motors


Experimental Testing (Kunz)


The detailed aerodynamic design of the rotors is highly dependent on the performance characteristics of a given motor. Manufacturers commonly provide some form of performance data, but it is often either insufficient, or based on theoretical models rather than detailed experimental data. In addition to the performance of the motors themselves, it is important to have some method for experimental evaluation of candidate rotor designs and validation of the design analysis tools.

Test apparatus to accomplish these tasks has been developed and is currently being fabricated. The test stand has the capability of evaluating rotor lift for all currently considered motors with the 3mm Smoovy motor representing the small end. Rotor torque should be accurately measurable for motors starting from the 5mm Smoovy. The torque output from the 3mm Smoovy is low enough that two problems appear. Bearing friction in the apparatus becomes an issue, adding uncertainty to the measurements. In addition, the test stand converts the torque to a force measured by a milligram accurate electronic balance. The small torque values require a very short moment arm to achieve a reasonable force output, but the minimum length is limited by the dimensions of the electronic balance. The result is that qualitative torque measurements are possible with the 3mm motors, but precision measurements currently pose a problem.

The evaluation of motor performance is accomplished by coupling the shafts of two motors. The one being tested is mounted to the test stand. The second motor acts as a torque load and is mounted in a fixture such that the drive shafts are co-linear. The motor being tested is powered by a constant voltage power supply. Torque, RPM, and current data are recorded as the load torque is gradually increased until the rated stall current is achieved. From this data, power and efficiency curves can then be calculated.

Motor Control Electronics (Fabian)

Discusion with company RMB, the manufacturer of the currently used motors and motor drivers, about the availability of smaller/lighter version of their existing motor driver. The dominant part of the motor driver is a 28 pin integrated circuit (IC). RMB disclosed the name of the IC. It is a Philips made TDI 5145. Samples of this IC were ordered. The effort on obtaining the silicon die of the TDI 5145 continues.

Battery System (Brunet)


At this point in time the project has adopted a strategy of continuing research on creating our own batteries, however, our main focus is finding commercially available batteries which can meet our requirements.
On the first point, SRI is making progress with the direct write process. They are working on integrating the full process. Their high power performance tests are quite encouraging, but at this stage they are building on a larger scale than we require.
The search for commercial batteries continues. Upon revision of our current information and testing, modification of testing procedures are considered in order to obtain more useful data for our specific purposes, in particular battery life.

Structures/Frame Development (Prinz, Cheng)

We are investigating approaches for an integrated frame system. One idea is to use a PC-board with the motor control and sensor components as the frame itself. We will try some experiments with this approach in the next month. We would also like to integrate some aerodynamic surfaces in the frame to enhance lift. These ducts may be molded over the frame.

Stability and Control (Fay)

A computer model of the four-bladed mesicopter is needed to analyze the dynamics of the vehicle. The goal is to determine whether the mesicopter can be passively stabilized, or if a closed-loop controller will be needed. To accomplish this, an initial 2D planar model has been created which allows for three degrees of freedom in horizontal and vertical translation as well as pitch. The aerodynamic forces of the rotors are modeled using classic large scale helicopter theory. Besides the mass and inertia, other parameters of the model can be changed including the location of the center of gravity and the tilt of the rotor shafts. The model shows that for low C.G. and inward tilting rotor shafts, the 2D model is passively stable. One problem with this model is that it is unclear whether large scale helicopter theory is valid at the scale of the mesicopter.

There are many possible areas for future work in stability and control. The rotor aerodynamics must be tested experimentally to determine whether our model is correct. A 2D linearized model is also being to developed to analyze the stability margins quantitatively. And eventually, a full-blown 3D model will be needed.

Sensors and Communications (Kroo, Holden)

We are continuing our evaluation of sensors for flight control and navigation as well as hardware for communications. Discussions with Integrinautics suggest a very interesting approach to cm-precision GPS-based navigation using an onboard system weight in the 1-g range. Developments in the area of very small packet-based RF tranceivers and web-server on a chip were forwarded by JPL researchers and we are looking into how these might be integrated in the future.

System and Application Issues (Kroo)

Discussions with researchers at various government labs continue. Ilan has briefed people at Langley, Ames, JPL, DARPA, and NAVSEA regarding recent developments. While our NIAC work focusses on 10-40 years out and mesicopters in the 3-10g class, most researchers interested in nearer-term applications would like 10-20g payload capability.



Edit

Last update: 11-Oct-99 3:06:40 PM
WebEdit servlet by I. Kroo