For the last 2-3 months I have been pursuing the possibility of evanescent waves being the source of the EM Drive effect. For those unfamiliar, http://en.wikipedia.org/wiki/Evanescent_wave
An evanescent wave is a near-field wave with an intensity that exhibits exponential decay without absorption as a function of the distance from the boundary at which the wave was formed. Evanescent waves are solutions of wave-equations, ...
I have some results.
Using FDTD software developed at MIT, and my 5 y/o home computer, I modeled the experiment done at NASA Eagleworks, presented here: http://www.libertariannews.org/wp-content/uploads/2014/07/AnomalousThrustProductionFromanRFTestDevice-BradyEtAl.pdf
I found the following salient points. First, my computer is not powerful enough to perform FDTD calculations in high resolution, so I'm stuck with moderate resolution at best. That rules out resolving the thin copper sheets on the circuit board ends of the cavity. Second, there isn't any data that I can discover describing copper at this frequency and the power level experienced by the cavity ends. Sure, lots of data on copper for shielding, but...
Consider that operating at Q = 22,000 and drive power of 2.6 watts with the cavity end diameter giving an area of 0.058 m^2 the radiation intensity approaches 1 MW/m^2 for the Eagleworks tests, and for higher power devices, 1 kW/cm^2. That is a lot of power and it seems no one has researched the behavior of copper under those intense conditions.
With those two points understood, I did calibrate the FDTD model to a photon rocket, giving thrust of very nearly 1/c as it should. This is with the detector plane about two cavity lengths behind the source antenna.
For the reasons given above, I used a perfect metal or ideal conductor as the material for the Eagleworks cavity model. Not surprisingly, when I ran the model totally enclosing the source with perfect metal, there was zero force detected. That result has been calculated analytically and discussed widely.
I went further, considering that the end is bolted onto the cavity, what if it leaks RF? I modeled a narrow slice around the circumference of the cavity cone in the end plate, a variable sized opening but at the smallest resolution my computer will allow, 0.2% of the cavity large end radius, about 0.27 millimeters for the Eagleworks device.
The simulation detected Force/Power ranging from 2/c to more than 3/c.
That is two to three times the thrust of an ideal photon rocket. I think the cause of the force is evanescent waves escaping through the very narrow gap in the cavity base.
My data does not answer the question, "What causes the thrust of the EM drive?" It does point in a direction.
Some may object that these results are no better than the Woodward device gives. It is not my purpose to create a new or different kind of thruster but to investigate the cause of EM drive forces. I do note that making a resonate cavity with a loose end cap should be much easier than making a Woodward device for the same thrust.
I'm sure someone will ask, so, “Yes, I did look at forces generated by similar gaps in the small end of the cavity.” Forces exist but only about one forth as great as at the large end. That may be directly due to the gap area, the small end circumference being about half that of the large end, and the dielectric does interfere. I should probably remove the dielectric and run some small end test cases. However, I suggest that a better solution when constructing a test article, is that the small end of the cavity be heavily plated, sealed tight with no leaks.
I can run more cases with low to moderate resolution and perfect metal, but as I wrote at the top, I need a bigger computer and better knowledge of copper behavior under intense radiation in order to calculate definitive results.