Do It Yourself Antigravity

The crucial test for any Mach Effect like device is in a full vacuum environment so here is my DIY vacuum test bed and proof that my piezoelectric (PzC) stack device does not work in high vacuum. I’ll describe the test bed, the device and the test results.

Vacuum Test Bed

The vacuum chamber is made from an inexpensive 6” x 24” PVC pipe called an SDR-35 riser pipe. It has two drive power wires passing through either side and sealed with E6000 glue. The pipe contains a wood frame holding two razor blades soldered to the drive wires. (The blades replace the guitar strings of Blog 8.) The balance arm, which contains the PzC stack, counterweight and needle axel, is balanced on the razor blades (Figure 1). This end provides access to the chamber and has the vacuum pump attachment. The gasket is made from a 1/16” rubber sheet affixed with E6000.

The two end pieces are made from a 1/2″ x 12″ x 12″ acrylic blank from Amazon. Figure 2 shows the viewing end of the chamber with end piece glued using E6000. The vacuum gauge is threaded into the pipe and sealed. The balance arm is seen with a 30 mg weight (a coil of nylon line) for the test at atmospheric pressure. That coil is removed for vacuum testing since this end of the arm weights 30 mg more at full vacuum. (Everything weights a little more in vacuum but Styrofoam is significantly heavier.) Testing at atmospheric pressure gave the same result as seen in Blog 8.

Figure 3 shows the chamber sealed with vacuum pump attached and ready for test except for connecting the drive wires. The pump is a Zeny, 5 Pa portable refrigerant tool (on Amazon USD 75). The piping can be almost anything that’s available. I used 3/8″ NPT and 1/4″ valves with a 1/4″ x 1″ nipple threaded into the end cap. The hose is fuel line clamped to air hose connectors.

Vacuum Testing

After removing the 30 mg rebalance weight I ran the pump up to 17″ vacuum and viewed the result of drive power on the balance arm. The resulting thrust was about 1/4 the thrust observed when the test is done under normal air pressure. Then the pump was run up to 29.5″ vacuum and the test repeated. At full vacuum there was no motion of the balance arm when the stack was energized. This is positive proof that my PzC stack does not generate any detectable force in vacuum.

I had previously attached a nose cone on the end opposite the forced direction and under normal air pressure found the same result as with no nose cone. That should have eliminated asymmetrical air resistance and that end was covered so micro jets were also ruled out. It may be that my device slightly lowers the air pressure at one end causing it to be pulled in that direction.

Conclusion

While my device is not showing results, Drs Woodward/Fern are reporting Mach Effect thrusts so I’m not giving up. Now that I have a way to test under full vacuum, there are a few ideas left to try and to look for a simple way to negate the false positive air pressure effect. I also want to thank N for the comment emphasizing the importance of the air pressure interaction. So, please, comment on any questions, error modes or suggestions that come to mind. Until next time, thanks for viewing. – Larry

While the salt water connected, VER 4.0, of my device worked well, it had repeatability problems of various kinds that are now addressed in Version 5.0. A later test proved my device does not work in vacuum (Blog 9). The thrust reported here is not the Mach Effect but probably due to some kind of air pressure interaction which I plan to investigate later. VER 5.0 is by far my best effort and is showing much improvement with repeatability. This will be the first of several installments regarding the VER 5.0 test stand and test results. Today the test stand will be briefly described, then a video will show vertical tests using piezoelectric chip (PzC) stack #7 previously tested in the salt water bath of VER 4.0.

Test Bed

The test bed was designed to allow 360 degree rotation of the stack in the vertical plane so that resulting thrust forces could be collected and compared in various orientations. It is made up of three parts: the balance arm, the support stand and the laser measurement device. The laser was previously described in Blogs 1 and 2 and will be used later. The stack is suspended with 4# nylon fishing line on the left end of the balance arm (Figure 1). The arm is made from 3/16″ Styrofoam poster board. The hub on the middle right is five layers of Styrofoam laminated with super glue and containing the support and contact needles. They are connected to the stack with 24 gauge silver wire for electrical contact. The counter weight on the right is Styrofoam and lead which can be removed for other testing modes. The complete balance arm is 55.9 cm long and weighing 15.3 g, is suspended on two powered guitar strings in the support stand.

The support stand is an A-frame construction of plywood and hardwood providing rigid support, alignment and connection for the guitar strings (Figure 2). The positive polarity string can be seen in the figure being tensioned with a rotating steel bolt. Both strings are tensioned to the same note in the upper range of the A string (~11 kg). They are mounted 55 cm above, aligned together and offset 3.75 cm from the plywood sides providing 5.0 cm separation. The offsets are hardwood, copper covered fret bars with electrical connection to the drive wires. The balance arm is shown hanging on the strings in the 90 degree orientation used for vertical thrust measurement. It was earlier tested in the zero degree orientation for horizontal thrust using the laser.

Vertical Measurement

A poster board with cm arc lengths is positioned behind the balance arm pointer so that movement can be recorded in video format. The first video shows stack #7 with the large mass pointing down. The led indicates when the drive power is energized with about 2 watts at 605 kHz (See Blog 7 for amplifier schematic). The thrust is up which is toward the small mass. (My design uses no added small mass; instead, it is assumed to be some unknown portion of the PzC itself.) One cm on the scale represents about 2 mg.

Another test was run with the large mass pointing up. The thrust is down as expected but it is only about half what was seen in the up video. This is an obvious bias problem. In my zeal to get this first vertical test completed, I carved the Styrofoam in a very nonsymmetrical fashion while stabilizing the balance arm. Regardless, these are really good results because it’s only necessary to average the two in order to remove the bias. The average thrust of both runs is estimated at about 15 uN. I also used a ~1.5 mg calibration weight ( 3.75 cm nylon line) to show the same approximate thrust effect as in the up and down orientations.

It was very surprising to find the large mass in stack #7 required only 17 mg which was just the square paper tube holding the PzC. This was discovered while attempting to determine the optimum weight for the large mass. (The naked PA2JEW chip used here weights 171 mg.) The paper tube can be bigger or a little smaller but when it’s a few mg the thrust disappears. Stack #4 reported in Blog 7 included a 170 mg brass large mass that worked well but at a lower frequency (444 kHz).

Conclusion

This is still an amateurish effort that could contain errors; however, I have made ever effort to eliminate false positives. For example, in order to go from the up to the down orientation I only need to rotate the balance arm 180 degrees. VER 5.0 test bed is an excellent way of testing PzC stacks in air but testing for the Mach Effect requires vacuum, which will be covered in the next post.

As before comment on any questions, error modes or suggestions that come to mind. Until next time, thanks for viewing. – Larry

From everything I can glean from the internet and with my amateurish results it appears the Woodward Effect is delivering a small “Mach like” force using only electrical current. I believe this can now be considered a new force in physics. Nothing has been published but there are reports of significant drive forces using Dr. Woodward’s MEGA (Mach Effect Gravity Assist) device in two new modes: Hanging from nylon line and moving on a virtually frictionless air bushing slide. MEGA is operating near resonance so the major force is toward the smaller mass. When a hanging device moves toward the smaller mass the so called Dean Drive argument is moot. (A Dean Drive device can move along a frictional surface in either direction due to the “slip-stick effect” but not suspended). A hanging Dean Drive can possibly move a very slight bit toward the heavy mass if the action of the vibrating small mass causes the device center of mass to shift slightly (on average) toward the small mass. There is no way a hanging Dean Drive can move toward the small mass nor can it move on a frictionless surface. The MEGA reported forces are still fairly small (3 mN) but 10^3 times more powerful than earlier and if it works in space and can be scaled, a spacecraft star drive will be possible.

So here’s your chance to work on a new force in physics. My latest Ver4.0 is easier to build than earlier versions and is delivering repeatable results in the direction of the smaller mass. But first I want to mention a controversy resurfaced with a lengthy paper by Dr. Jose Rodal in the very prestigious and mainstream journal of General Relativity and Gravitation, where he states in part, concerning the mass-energy fluctuations of Dr. Woodward’s theory:

In summary, this is a higher order effect and thus appears to be too small to be used in practical space travel application (unless the spacecraft is near a black hole or a neutron star).

Rodal, J.J.A.: A Machian wave effect in conformal, scalar–tensor gravitational theory. Gen. Relativ. Gravit. 51, 61–84 (2019)e

Dr. Woodward had previously not been published in mainstream theoretical physics journals, but has now responded to this criticism in the same journal and concluded in part about the Mach effect controversy:

The arguments, however, have been restricted to the general relativity community and “academic” until quite recently. For example, a decade and more ago, when the Higgs interaction was widely hailed as the “origin of mass” (mass being the meas­ure of inertia), no one even bothered to mention that Einstein took gravity to be the origin of inertia and that the topic had been long and at times hotly debated in the general relativity community. Rodal’s dismissive critique of Mach effects falls squarely within this tradition. A year or so ago, however, he sent me an apposite Feynman quote: “It doesn’t matter how beautiful your theory is. It doesn’t matter how smart you are. If it doesn’t agree with experiment, it’s wrong.”

General Relativity and Gravitation (2020) 52:11

I don’t understand enough general relativity to comment on this argument but it certainly appears that the MEGA drive is providing experimental data. Drs Woodward/Fern are getting great results, a few others have seen lesser results and even my DIY device is showing a few uNs of force. My Ver. 4.0 stack is lighter than previous stacks and now suspended over two saltwater reservoirs to provide the drive power. This design eliminates the hanging drive wires of the previous model that some argue could cause false positives. Achieving comparable results with the same stack using different drive methods is further confirmation that the force is real. The requested construction details follow or you can skip to the test results below.

Construction

Figure 15 shows the PA2JEW piezoelectric (PzC) stack with a .17g brass counterweight attached with 9340 Hysol epoxy. The drive wires were replaced with short pieces of 30 AWG copper with the strands separated (for high frequency isolation) and terminated with short pieces of 24 gauge silver wire (Figures 16 and 17). A square paper tube surrounds the stack and has the 2 lb test nylon fish line supports attached with superglue. The blue saltwater bath is made from a silicone ice cube tray with silver drive wire on the bottom. Two heaping tsp salt in 200 ml distilled water replaces the liquid metal previously attempted. These tiny devices are drastically effected by surface tension. All liquid metals have 10 times the surface tension of water. Also the water surface is affected by the ability of the water to wet the sides of the bath. Applying glycerin or similar to the bath sides provides a wettable surface that allows the bath to be more level. The laser and sled are described in earlier blogs. Drive power is supplied with an ordinary signal generator and with my simple amplifier design shown in Figure 18.

Test Results

Results below represent three tests showing mean values of 10 runs each. Stack #4 is driven at 444 kHz with a short pulse (0.5 s) that starts the stack swinging and followed by a long pulse (1.3 s) that holds it in an elevated position. Force is shown in blue, drive power in red and the error bars represent +/- 1.0 sigma about the mean. Because of the stacks small mass (~.34 g) it was necessary to reduce the power (< 1 watt) in order to keep the signal in the linear range. Figure 19 shows results from the laser measuring movement from the east end of the stack which is also the large mass end. Figure 20 shows results from the laser measuring movement from the west end of the stack which is the small mass end. The stack had to be repositioned in the west test resulting in an small increase in power and force. Comparing the two test reveals a clear thrust of about 2 uN toward the small mass. The stack was then reversed 180 degrees and the last test ran again. This resulted in Figure 21 which is still measuring on the west end and now the large mass end. Notice there is a small tendency to diminish the positive laser measurements (bias) probably because of nonlinearities. Normally the last result is subtracted from the previous one thereby erasing any such systemic bias. This results in Figure 22 which clearly shows a few uNs force toward the small mass.

Variability from one test to the next is less of a problem in Ver. 4.0 but is still an issue. The Ver. 4.0 stack has much less mass and only one epoxy joint so those factors likely allow the stack to resonate at 444 kHz which is near the PA2JEW design frequency (450 kHz). Previous versions had much larger mass, nuts, bolts and two epoxied joints that likely required the higher frequency (616 kHz) to fluctuate the mass. The higher frequency of Ver. 3.1, being close to 3/2 the fundamental, always delivered forces toward the large mass (Notice the PzC length/width = 3/2). Now with Ver. 4.0 operating near the first harmonic, the force is toward the small mass.

This Mach effect force reversal was predicted by Dr. Rodal , although at twice the fundamental, where he modeled the PzC stack using a partial differential equation. He obtained an exact continuous solution, one result of which is shown on Slide 60 of his presentation at the 2017 Advanced Propulsion Workshop. What appears to be a theoretical reversal in his recent paper, but isn’t, may be explained by his comment concerning a 10^2 or 10^6 scale factor required for Slide 60. Regardless of controversies and scale factors, with my amateurish results and primarily with the laboratory grade thrust reversal results by Drs Woodward/Fern, this is powerful confirmation of the Mach effect.

As before  let me know any questions, error modes or suggestions that come to mind. Until next time, thanks for viewing. – Larry

After reading the Technische Universität Dresden (TUD) report and realizing that the on/off power cycle for the California State University, Fullerton (CSUF) thrust devices (MEGA) were not responding according to predictions (Blog 5), I decided to look for switching transients on my DIY Mach Effect Device (DMED). I should also mention that the MEGA on/off power cycles were behaving according to predictions at CSUF as reported in the NASA document. Dr. Woodward stated,“This follows immediately from the dP/dt term in the mass fluctuation equation (Equation 8).” After shifting 90 degrees TUD reported switching transient thrust in the orthogonal direction which does not agree with CSFU results. Just as a reminder, that failed test at TUD appears to have passed with DMED since my stack always swings in a fashion representing unidirectional force and nothing appears in the orthogonal direction. But there is much more to this switching transient story. Is it possible that I’m only seeing on/off drive forces and seeing nothing during the steady state drive portion? I don’t have a torsion balance machine to measure forces directly during the drive cycle. So, I have to determine thrust forces indirectly by measuring the resulting motion on my hanging stack. This next test will vary the drive pulse duration to see how that effects the stack motion.

To determine how much thrust was applied to the stack, I measured the range of motion under varying drive durations. A large “power on” impulse would be more obvious for short drive times unless the “power off” sequence erased the effect because it would be opposite in polarity. That positive “power on”, negative “power off” effect was evident in CSUF and TUD data. The first plot, Figure 13, shows the effect of increasing the drive length from 25 to 100 ms in 25 ms steps. The resulting negative motion at 1.9 s increases normally. (As always in my data, drive is toward the large mass.) The motion at different drive durations appears also to be approximately linear with no constant offset. (within +/- 1 sigma). Therefore, there was no large impulse. This same result would also be evident if the “turn off” sequence erased the “turn on” sequence. In order to eliminate that possibility I looked at longer drive cycles that ended before and after the hanging stack starts back to zero, sometimes called the “turnaround “ point.

In the next two tests shown in Figure 14, the “power off” points bracket the turnaround point which is ~470 ms after “power on”. The before turnaround “power off” test is at 430 ms and the after turnaround “power off” test is at 530 ms. If there is a significant “power off” thrust and it is opposite to the “power on” thrust, it would provide opposing movement force before turnaround and aid movement force after turnaround. In that case there would be a large difference in the positive extent of the stack on the before and after tests. An can be seen around 2.4 s, there is very little difference in the before and after turnaround tests; therefore, there is little or no “power off” impulse. The last point to nail this issue is the fact that the stack response increased so dramatically from Figure 13 to Figure 14 by simply increasing the drive time regardless of any supposed switching transient.

So, there is no measurable switching transient in DMED, but why not, if CSUF always measures this force. CSUF has a much lower drive frequency (35 kHz) than DMED (616 kHz) and I’m fairly sure their power on/off sequence is much faster than my device produces. That could generate a significant change in power (dP/dt in Dr. Woodward’s equation). I was unable to find a good way to switch the signal generator so I merely switch the amplifier power on/off with a solid state relay. This is a fairly slow process compared to my 616 kHz frequency transitions; therefore, there should be very little change in power .

DMED continues to agree with the Mach Effect equations and I will continue to look for ways that disagree. So far my tests agree but I haven’t tried everything. Later I’ll show that the force is toward the small mass when the device can operate close to the PzC resonance. As before  let me know other error modes, questions or suggestions that come to mind. Until next time, thanks for viewing. – Larry

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.

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

Testing for false positives on my DIY Mach Effect device (DMED) is important because it is possible there is some other phenomenon causing this apparent Mach Effect force. By reversing all the configuration constraints it will become obvious because the force will switch with the problem should it exist. The plots in the previous blog used a stack with the large mass having positive polarity. To eliminate polarity and direction, I switched to a stack with negative polarity on the large mass end and rotated the stack 180 degrees. (Later I’ll show the result of doing only the rotation.)

Figure 10 shows the results for this example, still measuring on the west end with laser pointing north. Response, in blue, is the average of 50 runs and the drive voltage is in red. As before, when first energized, the response drops, this time to -40 mV. This is not thrust, it is mainly caused by crosstalk in the PicoScope which is now corrected.(This effect is confirmed in Figure 11). When the PzC expands the stack center of mass shifts very slightly toward the PzC causing the stack to rebalance very slightly toward the large mass end. Since I’m now measuring on the large mass end, that causes the photo detector output to drop very slightly. That effect is true but the major portion of the change was due the crosstalk. The opposite happens when the power is turned off and the PzC collapses to its original dimension. The crosstalk does not minimize what happened between those two events when the stack is vibrating at 616 kHz. The response went down from -40mV to -58 mV between 1.4 s and 2.0 s. That -18 mV toward the large mass is real thrust because the stack immediately begins oscillating. The thrust is less than in Figure 8 and the sigma’s are larger. There is considerable variation from one test to next and each stack has a different response curve; however, the forces always remain consistent regardless of the change. One of my goals for the future is to reduce this variation.

In order to confirm that the PzC expansion does not cause thrust, the previous test was run again with the signal generator de-energized. Figure 11 is the averaged result of 50 runs showing the -40 mV change for the duration of the drive with DC voltage only. After the drive de-energizes the stack is barely moving for the remaining ~4.0 seconds. The oscillation is now less than 1.0 mV p-p and when compared with the 27 mV p-p oscillation of Figure 10, the crosstalk problem, the PzC expansion and collapse is not causing significant thrust (With the crosstalk corrected these tests will be rerun later.) Later I’ll show that the force is toward the small mass when the device can operate close to the PzC resonance.

This continues to suggesting that the Mach Effect is actually driving the stack or there is some other unknown something causing the stack to swing so consistently. As before  let me know other error modes or suggestions that come to mind. Until next time, thanks for viewing. – Larry

The reversible stack worked very well however, the power and support wires were effected by the heat generated in the piezoelectric chip (PzC). I was powering the stack for 5 s and heat would travel from the PzC to the first brass mass, then to the nickel bronze guitar string attached to that mass. The same would happen to the second brass mass later. The first string would expand faster than the second string causing a bias that would look like thrust toward the large mass. Even worse, when the stack was reversed everything would reverse and still the thrust appeared toward the large mass. The big clue was that the thrust did not release immediately when the power was cut, but slowly diminished. The problem was easy to verify. I just pinched the first string with wire cutters pliers just above the stack. Rerunning the test showed none of the original thrust. Having power and support with the same wires was a great idea that didn’t work. Their still appeared to be thrust so I went next to Build 3.0 (See Figure 7)

I wanted to reduce the dampening the the stiff guitar strings caused so I switched to 4 lb test nylon fishing line. It’s really like thread but, also being nylon conducts heat very slowly. I also increased the string length to 25 cm to reduce the gravitational force on the stack. Now called the DIY Mach Effect Device (DMED), Figure 7 shows the other significant change which was attaching 30 awg stranded copper wires directly to the top of each support screw. I also separated the strands ~ 3 cm to allow the stack to vibrate freely while simultaneously providing power.

I started running data with a very short 0.5 s, 4 watt power cycle to reduce the PzC heating. I got great results as you can see in Figure 8. When first energized the response drops to -53 mV but this is not thrust. (That was confirmed later by turning off the signal generator and driving the stack with 13 VDC causing no stack oscillation.) When the power hits the PzC it immediately expands  ~0.3 um (according to Thorlabs). Since I’m measuring on the PzC end that slight expansion causes the photo detector output to drop slightly. That’s not what happened here as I later found. The lions portion was due to crosstalk in the PicoScope which is now corrected. I will redo those runs in a later blog. That problem does not minimize what happened during the period when the stack is vibrating at 616 kHz. The response went from up from -53 mV to -26 mV between 0.9 s and 1.25 s. That 27 mV toward the large mass is real thrust because the stack immediately begins oscillating. The period is easy to calculate in Figure 8 at ~.9 s.

With that result I decided to try multiple drive cycles at the natural stack period .9s. Figure 9 shows the result of 5 drive cycles where the stack is driven to 230 mV p-p. The obvious next step was to drive for 30 cycles and video the motion (The problem with the PicoScope is not a factor here since it’s not in the circuit). I placed an led in the frame so you can see when the stack is energized.  It weighs about 2 g and is moving about 1 mm p-p after 20 drive cycles which means it is responding to about 2 uN force on each drive cycle. If you look carefully you can see the thrust is always toward the large mass. If confirmed by Dr. Woodward and others, the Mach Effect will be a new force in physics or maybe a new gravitational force since they are using Einstein’s field equations.

In order to eliminate false positives, I tried different orientations of the stack: reversing, changing polarity, different compass directions and using a different stack. All caused thrust toward the large mass. Later I’ll show that the force is toward the small mass when the device can operate close to the PzC resonance. I also ran another video with just the PzC and no large mass to show there was no drive thrust.

This is all suggesting that the Mach Effect is actually driving the stack or there is some other unknown something causing the stack to swing so consistently. As before  let me know other error modes or suggestions that come to mind. Until next time, thanks for viewing. – Larry

Having witnessed some small photo detector changes reported in the previous blog it is possible I am measuring some small Mach Effect forces when the piezoelectric stack is energized. The obvious requirement though is to eliminate false positive and misleading data. The objective for Build 2.0 is to develop a completely reversible stack so that apparent thrust can be measured on subsequent runs and common mode errors subtracted out.

Figure 4. shows the overall Build 2.0 design from below with reversible stack (negative stack). The brass counter weight is now divided into two pieces and electrically isolated from each other so that polarity sensitive power can be brought down via the support wires. In the figure the uppermost mass is mechanically attached to the piezoelectric stack and electrically connected to the negative, right side of the piezoelectric stack. The middle mass is electrically connected to the positive, left side of the stack. Each brass mass has a hole drilled to accept brass screws soldered to the power/support wires. This design allows the entire stack to be unscrewed and reversed 180 degrees between data collection runs. This test is called the Orientation Reversal test. As you might guess this reversal or any change to the stack disrupts alignment meaning the laser must be readjusted.

Figure 5. shows the Build 2.0 laser sled which addresses the adjustment issue and the alignment problem in general. Laser and photo detector are aligned and mounted on an aluminum plate that moves back and forth on nylon slides mounted on a second aluminum plate which slides independently left and right. The adjustment screws have a thread pitch 0.5 mm allowing very accurate alignment with the stack in two dimensions, forward/backward and left/right. The third dimension, up/down is adjusted by moving the pendulum support with nuts/screws. Figure 6. is a close-up of the infrared detector used to measure the piezoelectric chip temperature. Since all power runs are at the anti-resonance point (~600 kHz) heating is very rapid and must be monitored carefully. These piezoelectric chips loose their polarity and their usefulness beyond 260 degrees Fahrenheit.

Figure 6. also shows the negative stack directly above the IR detector. It is electrically connected so the large mass end has negative polarity. This means that the electrical connection above the supporting wires must agree. When the stack is rotated 180 degrees the connection above must also be switched 180 degrees to maintain correct stack polarity. The support wires are not moved therefor the current flow before and after the switch is also reversed. To eliminate all false positives we want to keep everything constant before and after the switch. To eliminate the current flow issue we build another stack (positive stack) identical in design but with the piezo stack electrically connected to the large mass end with positive polarity. For this test we collect data with the positive stack, replace the positive stack with the negative stack and collect new data. Differencing the sets removes any common mode errors due to current flow direction. This test is called the Stack Swap test. Results of these tests suggest small Mach Effect forces and will be detailed in the next blog. As before  let me know other error modes or suggestions that come to mind. Until next time, thanks for viewing. – Larry

What if there was a device that could defy gravity and lift a refrigerator off the floor so anyone could push it out the door. As fantastic as that sounds and as unlikely that it will actually work, it is one implication of the Mach Effect Thruster. If it does work and could be scaled up it would revolutionize just about every industry in the world. It would cause changes more than semiconductors had done and more on the scale of the Industrial Revolution. I want to build and test a very small Mach Effect antigravity device at home.

Is It Antigravity?

The physicists who have been working on the Mach Effect Thruster never use the word antigravity probably because it sounds a bit dramatic. If it works it will be able to provide enough thrust to overcome gravitational fields. Also I’m not a physicist concerned about reputation so I’m calling it antigravity.

It’s actually called the Mach-effect gravitational assist (MEGA) drive, invented by James F. Woodward , a Ph. D. physicist currently at California State University, Fullerton. All current spacecraft carry lots of propellant and expel it out the back in order to generate forward thrust. The MEGA drive purports reaction-less thrust using only electricity for propellant. Dr. Woodward and a few others have published laboratory grade work showing small thrust results around 2 or 3 micro newtons (uN). A recent NASA document reports measurements of 100 uN thrust. A lessor number of equally capable researchers have published results of similar devices that can not produce 1 uN thrust. Nevertheless NASA has seen enough to continue funding Dr. Woodward’s work currently to 500,000 USD. A definitive test of the device is currently scheduled at the U. S. Naval Research Laboratory in Annapolis, MD notwithstanding Covid complications.

Most physicists remain skeptical primarily because MEGA relies heavily on Mach’s principal which they believe violates general relativity. Mach’s principal is commonly stated as in Wikipedia, “local physical laws are determined by the large-scale structure of the universe” or parochially “mass out there defines inertia here”. There are many Machian gravitation theory formulations; but the two most prominent are, Hoyle–Narlikar and Brans–Dicke . Experimental data from the Cassini probe appears to invalidate the Brans-Dicke formulation. These two formulations have been used to derive equations that predict the forces supposedly generated by the MEGA device. With the exception of this MEGA data I believe there is no experimental evidence to support Machian gravitation theory.

My goal with MEGA is to build an inexpensive workbench device that will provide evidence albeit amateur evidence to either support or contradict MEGA theory. What makes me think I can do this when much more capable researchers have had such differing results? All published work uses devices that weigh around 150 g and operate around 35 kHZ. I am working above 600 kHz with about 1-5 g devices. MEGA mass effect gravitation equations suggest frequency has a rather large exponential scaling factor. Even if it is only frequency squared which is at the low end of the mathematically predicted scaling effect, 600 kHz will provide ~300 times thrust multiplier. Using a much smaller mass may or may not help. Considering everything I hope to produce ~100 uN thrust which can be measured on my workbench. My interest include electronics, micro computers and amateur physics so this effort fits to a T.

Design

I plan to document everything on this blog and provide enough information so that other enthusiasts can built on this effort, hence DIY antigravity. The small results that I see to date need to be scrutinized and tested. Feel free to question everything and comments are also welcome. Why are reputable physicists using larger devices? It may be because high frequencies require small devices which raise a host of other problems. This may be a quixotic venture but for me it is entertainment and if I can produce consistent results whether positive or negative it is worthwhile. My device is inexpensive and designed so it can be reproduced by any skilled amateur. All components for the device shown and tested below cost about 200 USD.

There have been lots of devices tried and rejected for various reasons over the past year so I will ignore the experiments that failed to give consistent data. My system design is shown in block form in Figure 1. The oscillator is an inexpensive signal generator which was convenient for my testing. It isn’t required since the signal is narrow band and can easily be produced with a simple breadboard adjustable oscillator. My design for the amplifier is easy to build and provides the stack drive signal. It must produce about 20 watts of power at ~600 kHz and all positive (0 – 40 volts) because the piezoelectric stack cannot tolerate negative voltages. The forward/reverse power sensor measures power going to and reflected back from the device. Although not absolutely necessary it has been helpful to match impedances and improve resonant frequency selection. The device stack consists of one or more piezoelectric stacks glued together with a brass mass counter weight and suspended with a 25-250 mm thread to form a pendulum. The thrust detector consists of a 6 USD laser and 10 USD photo detector aligned with the end of the device. This provides a measure of movement from the vertical plane thereby implying thrust. Dr. Woodward and others are measuring thrust directly with sophisticated balance devices way beyond my means. The last block represents computer control to all blocks and data collection from the blocks.

Very Preliminary Results

Each block will be covered in great detail later but for this first blog I wanted to report some very rough and preliminary results using a subset of Figure 1. Figure 2. shows my first successful build of a thrust measurement device using a pendulum and laser apparatus as viewed from below. The laser on the left emits a beam that strikes the end of the device stack suspended right beside it so that only a small piece of light passes by and is captured by the photo detector on the right. The stack consists of a 9mm long 1/8″ square brass mass glued to a single piezoelectric stack. Since the detector is about 20 cm distant any slight movement of the stack will cause a large change at the detector.

Figure 3. Shows this test device with mounted laser and stack hanging by polyester thread. The photo detector is outside the frame with its output currently routed to an oscilloscope to facilitate tuning and measurement. Since mean DC voltage is being measured I believe this could also be accomplished with a voltmeter. With the laser on I adjusted everything so that about 300 mV was showing on the scope with the drive power off. After the stack is energized the scope output increases by 20-30mV to ~325mV. When the measurement pair is shifted to the back side of the device and readjusted to 300mv the drive power causes the output to decreases to about 275 mV. It appears the stack is not just expanding but is actually moving. This suggests there is a small thrust pushing the stack against gravity and friction in the direction of the brass mass. How much thrust this measurement actually represents will be addressed later.

Eliminating False Positives

Obviously this thrust could be caused by something other than the MEGA force. I’ve thought of three possibilities, electromagnetic effects from the drive signal, the local magnetic field or asymmetric resistance in the pendulum. The first one happened before on a previous design using a load cell so I am skeptical. In order to test this I hung a matching piezoelectric stack and wires down alongside the device shown in Figure 2. I then drove this second chip at resonance frequency with the same power previously used. There was no measured motion of this stack theoretically because it had no brass mass counterweight required in Dr. Woodward’s equations. Looking for magnetic anomalies I rotated the apparatus 90 degrees and 180 degrees and repeated the Mach-effect test. The results were consistent with the first test.

While this is interesting data the apparatus is touchy and too unstable to perform automated tests so it will be torn down. The next build will be more stable and must include a method to eliminate asymmetrical resistance bias. Quoting Dr. Woodward, “Direction reversal is a critical test for false positive thrust signals.” That test requires two runs with the the stack rotated 180 degrees for the second run with everything else remaining the same. This turns out to be very difficult for my design but I’ll be working on that. Meanwhile let me know other error modes or suggestions that come to mind. Until next time, thanks for the viewing and be safe – Larry