I just reviewed an interesting paper from the exceptional laboratory at the Institute of Aerospace Engineering, Technische Universität Dresden (TUD), where the authors completed a thorough analysis of the Woodward’s MEGA Drive and reported negative results. I am grateful one of the authors, Prof. Dr. Martin Tajmar, forwarded a copy via ResearchGate. (FYI ResearchGate is a great way to add cred to your “citizen scientist” status.) Aside from a large switching transient, the only tiny force they observed (~0.4 uN) with a very sensitive torsion balance using the MEGA Drive in a vacuum chamber was clearly not due to the Mach Effect. Moreover, neither the small steady state nor the switching transient force changed when the MEGA stack was rotated 90 degrees. Quoting from the paper, ” Hence, the effects observed by Woodward using the MEGA drive on a torsion balance can be explained by thermal and vibrational artefacts using Newtonian mechanics.”

On the positive results side, I keep testing my DIY Mach Effect device (DMED) every way I can imagine and continue to get consistent results although not always agreeing with the Woodward equations. I get good agreement using different drive voltages, large/small mass ratios, orientations, directions, polarity and changing devices but I don’t think frequency responds as predicted. DMED works best near the anti-resonant frequency and not much otherwise. I’ve independently measured the stack action at different frequencies and that’s where the action is located so it should be most effective at anti-resonance. I’ll do some runs at different frequencies later and report them, but for now, I want to show a recent result correcting the cross talk problem mentioned in my third blog.

Figure 12. Fifty point average response for DMED with ~0.5s, 616 kHz drive. The stack has negative polarity on the large mass end, which is facing east and measured on the east end with laser pointing north . Error bars represent +/- 1.0 sigma.

I’m calling this one Build 3.1 since I added a MCP6001 buffer to the Channel 1 signal to the PicoLab and switched back to the Forward Power for channel 2. The higher drive voltage used before on Channel 2 was causing most of the cross talk and the buffer matches the PicoLab input impedance which has helped reduce noise sigma. This test again uses a stack with large mass having negative polarity as in Figure 10 (Blog 4) except now the internal polarity of the piezoelectric chip (PzC) is reversed by simply gluing the large mass onto the other face of the PzC. I have been assured by ThorLabs that all PA2JEW chips are laid up and marked identically so it is easy to identify the left face from the right face. Figure 12 shows the results for this example, measuring on the east end with laser pointing north. Response, in blue, is the average of 50 runs and the drive power (~2 watts) is in red. With Build 3.1 and this ~ 2.0 g stack the 16 mV response, I’m guessing, represents about 1.5 uN and notice that the sigmas are better. Later I’ll show that the force is toward the small mass when the device can operate close to the PzC resonance.

The negative result from TUD don’t necessarily condemn the Mach Effect or the MEGA Device. The MEGA Stack is assembled by hand, it contains piezoelectric material that can degrade, the torsion balance has it own issues and measuring uNs is always difficult. My positive results are also not conclusive but they continue to suggest that the Mach Effect is actually driving the stack or there is some other unknown something causing the stack to swing in a Mach Effect fashion. As before  let me know other error modes or suggestions that come to mind. Until next time, thanks for viewing. – Larry