E. Flaws in the Resonant Cylinder and the Associated Design Iterations
0 likes140 views
This document summarizes design iterations made to address arcing issues in a resonant cylindrical cavity used in an experiment. Small gaps between components led to arcing when high power was applied. Design changes focused on grounding the movable metal end plate and connecting the waveguide adapter to the cylinder. Nylon screws caused arcing and were replaced with a press-fit plate attaching dielectric material. Copper mesh and conductive putty were used to improve grounding between components, but could not fully eliminate arcing without a custom-designed electromagnetic choke or precision waveguide matching. Further testing was postponed due to budget and manufacturing limitations for precision microwave components.
E. Flaws in the Resonant Cylinder and the Associated Design Iterations
- 1. Flaws in the Resonant Cylinder and the Associated Design Iterations California Polytechnic State University, San Luis Obispo Aerospace Engineering Department By: Brian Kraft and Kurt Zeller After completing the manufacturing phase of the experiment we attempted to connect the cylindrical cavity to our pendulum and execute test runs. This proved to be a more complex issue then initially thought and an educational experience in waveguide manufacturing and operations. Unfortunately, it proved to be extremely difficult to provide power to our cavity without inducing arcing at one or more of the metal interfaces. Grounding different components of the cavity became a primary focus, resulting in a number of new designs. The different configurations and the motivation behind each one will be described in this document. I. Introduction Understanding arcing and the potential sources of arcing was the first step in mitigating this complex issue. Research proved that arcing within waveguides is greatly based upon the frequency and power of the source that is used. Working at 2.45 GHz meant that our signal had a wavelength of about 12 cm, therefore any perturbations in geometry on the order of a cm had a significant impact on the propagation of electromagnetic waves. This trend becomes even more exaggerated as higher frequencies. Although this is an important consideration it was likely the high power magnetron which induced the majority of our arcing. Small gaps between components on the order of millmeters proved to be an issue when high power was used. These small gaps prevented uniform charging of adjacent metallic components which inevitably led to arcing at multiple points throughout the cylinder. As the cylinder design evolved, these changes were centralized around the sliding metal end plate and the connection between the waveguide adapter and the cylinder. II. Location of Arcing After each test run was conducted the cylindrical cavity was opened to observe the location of arcing. The high power that was used made this a fairly simple task and arcing could typically be located within a minute of decomposing the cavity to its individual parts. The first source of arcing observed in this experiment occurred between the nylon screws that were used to mount the movable plate and the walls of the interior of the cavity. After conducting longer tests runs, it was observed that the energy would be dissipated through the nylon screws to the metal plate located at the backend of the movable plate. These longer duration runs resulted in partial melting of the screws and forced a new design or different materials for attaching the HDPE section to the movable plate. In hindsight, using nylon screws was poor decision because nylon is a polar material and therefore experiences intense heating when bombarded with high amount of microwave power. The specific changes in design that resulted will be discussed in the following section. Another source of arcing was between the edges of the movable plate and the inner surfaces of the cylindrical cavity. Small black singes were observed within the cavity and corresponded with marks
- 2. around the diameter of the movable plate or the screw used to fix the movable plate to the shaft forming the plunger. This led to design iterations that attempted to uniformly ground the movable plate to the walls of the cavity and different diameter plates for a smaller contact distance. Throughout the experiment arcing was also observed between the connection of the microwave waveguide adapter to the cylindrical cavity. The microwave waveguide was removed from one the same microwave that the magnetron was obtained from and therefore was matched well with that component. In order to remove this waveguide a band saw was used to cut through the thick metal surfaces obtain the waveguide independent of excess metal. Operation of the saw resulted in a small, wavelike edge due to the compression and shearing involved in this cut. Attaching this heterogeneous surface to the flat surface of the end of the cylinder resulted in asymmetric grounding and small gaps where arcing was observed. Furthermore, it was later discovered that a nonconducting layer of paint had been applied to the surface of the waveguide during manufacturing. Eliminating arcing became the greatest setback in this project and many weeks were spent attempting to fix this issue. These attempts are outlined with great detail in the following section. III. Design Iterations In order to prevent arcing to the nylon screws either new screws had to be used or a different end plate needed to be manufactured. New screws were briefly considered but it was determined that manufacturing screws entirely of HDPE was not worth the time or raw materials. One brief design iteration involved countersinking the nylon screws halfway through the HDPE plate and placing small HDPE plugs overtop of the countersunk areas. The logic behind this design was that arcing might be less likely to occur to the screws if the electromagnetic wave front met a uniform section of HDPE rather than a plate of HDPE with two protruding screw heads. Unfortunately this did not mitigate the issue and the screws continued to be arced to resulting in rapid heating. Therefore a new methodology need to be devised for fastening the HDPE plate to the plunger without the use of any type of screw. This resulted in the manufacturing of a press fit plunger where a piece of HDPE could simply be pressed into the metal plunger. Metal screws were used to attach the metal press fit component to the backend of the movable plunger. These screws were countersunk within the plate so the electromagnetic waves would travel through a uniform HDPE surface before hitting a uniform aluminum surface. The edges of the press fit plate were angled slightly inward to allow a strong grip on the piece of HDPE that was used. This solution effectively eliminated arcing to the surface of our dielectric (the HDPE). Prior to implementing the pressfit plate a number of different plates were used to provide a conductive interface between the interior of the cylinder and the movable plunger. One technique involved fastening two grounding wires along the screws of the movable plate and connecting them to the bracketry required to hold the plunger. After this solution failed aluminum foil was wrapped along the diameter of the plate in hopes that it would compress as needed to provide flush contact to the interior of the cylinder. A third solution involved machining a closer diameter aluminum plate which contacted the interior surface at a number of different points. The asymmetric contact proved to be an issue and once more this design failed. The final attempt made to improve the movable plunger involved cutting out a number of circles of copper mesh that could be fastened to the movable plate using the two mounting screws. This mesh protruded beyond the diameter of the plunger slid along the edges of the interior as the piece moved. Eliminating arcing between the waveguide and the cylinder required several different designs. After discovering the tell tales black singes, the edges of the waveguide were hammered and trimmed properly allow a more uniform contact area. Although this method did not succeed through the process of tampering with the waveguide it was learned that the exterior surface paint was nonconductive. A grinder was then used to remove this paint but unfortunately arcing persisted. The next idea that was implemented was purchasing Ox-Gard and smothering the edge of the cylinder with this conductive putty. Ox-Gard proved to be extremely messy and although it reduced the perceived severity of the arcing it did not entirely eliminate this issue. Eventually a very fine copper mesh was purchased and layered appropriately
- 3. along the exterior of the cylinder. When the waveguide was fastened to this interface the copper sheets would compress allowing full contact along the surface (although the degree of compression along the diameter was not consistent). IV. Conclusion and Final Design Considerations Although many of these ideas seemed promising at the time they all eventually failed to prevent arching. At this point in time it is difficult to determine what the true issue is that is causing arcing. Many of these design iterations prevented arcing at a specific spot but failed to prevent this phenomenon as a whole. The severity of arcing has been drastically reduced with the implementation of each new design but it cannot be entirely eliminated. After a small initial arc the cavity does not continue to arc or arc again throughout the course of an approximately 20 second test run. Rather than continuing to fight this issue a decision was made to delay testing until the truncated cone was depleted. Had the team been motivated to continue pursuing this geometry there are a number of changes that could be made which reflect the knowledge gained through experimentation. The most obvious change that would be made is the implementation of an electromagnetic choke. This devices would have required a number of different layers or a complexly machined part to implement. Figure ____ shows an example of an electromagnetic choke. Because the mode seen in our cylinder is circular, the slots shown in the figure would need to extend along the entire circumference of the plate. These slots allow any electromagnetic energy that might be escaping from the cavity to be absorbed therefore preventing a buildup of surface current. An additional change that could have been made is the design of a specific waveguide for input to the cylinder. EMPro could have been utilized prior to building the waveguide to determine the functionality of this component with our source. After constructing this piece, it could then be brought to the VNA to ensure it is well matched with an actual magnetron antenna. This same methodology could be applied to the cylindrical cavity that was used in this experiment. The cavity was constructed prior to determining how to use EMPro therefore it was not based off a specific simulated result. Only the theoretical formulas were used to determine the appropriate dimensions of the cavity. The combination of these two designs would likely allowed for higher quality resonances and the propagation of TE modes rather than TM modes which are theorized to be better for thrust production. Although there are a number of ideas that could be implemented, this project is severely limited by budget considerations and manufacturing. If these waveguides and components could be purchased from companies such as Gurling Electronics one could assure that arcing would not occur and the signal would have very low reflectance. These off the shelf waveguides are extruded and often designed to customer specifications therefore making them extremely expensive pieces of equipment. A simple rectangular waveguide costs close $1000 which was beyond the scope of the budget for this project. On the other hand manufacturing these components would require large amounts of raw aluminum or copper which
- 4. ultimately would be very expensive. Even after purchasing this material there is still a likelihood that arcing could occur due to the precision of the CNCs we have available on campus. There are CNCs that can machine with nanometer precision which would ideal for microwave applications. After speaking with our industry expert, John Gurling, it became clear that microwave engineering is a great deal of guess and check. Even after a design has been produced it must still be tested with high power or high frequency electronics and occasionally these parts do night align properly. Most microwave systems feature tunable components to absorb reflected power or ensure resonance at specific frequencies. The evolution of the plunger described earlier in this experiment proved how difficult these tunable components can be without the proper manufacturing capabilities. John described microwave engineering as a field that involves a bit of “black magic” and sadly, we are not magicians.