Mach Effect progress

[GIThruster]

I'm thinking about the mass equivalent of the energy in the capacitor. This being pushed back and forth one way would produce a signal in the balance different from pushing it in a different direction.

Apart from the fact that the mass/energy you're talking about is too small to measure, it is being pushed back and forth 40,0000X/second, so a balance with a 7 second period is not going to measure what you're talking about. The time averaged force from AC electron motion is zero.

Looking at a photo of the device, the feed blatantly violates what is practically rule One of RF design: keep the area enclosed by circulating current as small as practical. As such, the feed may be expected to induce currents in any nearby conductors, and be attracted to any nearby ferromagnetic materials.

All of the feeds are shielded. Not sure what pic you're looking at but it's probably one of the thruster outside its Faraday shield, or with the shield cover off. While on the balance, the thruster is in a Mu-Metal lined, grounded steel box. BTW, both Paul March and the PhD from Oxford, Duncan Cummins; had early objections to the shielding on the balance. Jim made corrections based on that input in 2006. The guys keep tabs on what Jim does and he is very humble in response.

[Diogenes]

Which brings up one of the challenges. If the circuit is not in electrical resonance, the voltage and current are out of phase and the power into the cap is just not there at the right time to produce M-E. So it needs to be Z-Matched. If the M-E generation and rectification are not in phase, there can be M-E but no thrust. Fortunately in the current design, the 1w signal that generates the 2w M-E, and the 2w rectification are locked into phase because it is the PZT that creates that phase relationship.

If the resonant frequency isn't known within a narrow boundary, direct RF drive is probably not feasible.



I'm also wondering if perhaps one of the issues is that the natural resonance of the PZT is different from that of the Capacitive circuit being used. According to theory, one of them HAS to be driven at the first harmonic. To get this to work together naturally, the resonance of the one of them must closely match 2X the resonance of the fundamental for the other.


And of course, you have the physical coupling to contend with as well. I can see where this might get difficult to tease a thrust signal out of it. A lot of variables have to balance out.




As a possible different approach to powering this thing, RF coupling could still be used, but the coupling frequency could be way above the fundamental drive frequencies, converted to DC, and then converted back to the necessary drive frequencies and phases.


This would require a lot more circuitry, but may in fact be easier than dealing with all the cross coupling, resonance, impedance matching, etc. from using direct RF drive of the elements. One issue will still be how to develop a practical heat sink for KW level drive systems while operating in a vacuum.

Might be able to build something, but it might not be able to operate for longer than a few moments. (Same problem the polywell had during experiments. )

Cooling systems that work in vacuum require large surfaces to radiate heat. Not really practical for a small system on a balance. This is not so much of a problem with direct RF drive, because everything but the load would be external, and therefore would be able to be cooled by the ambient.


Interesting problem.

[Diogenes]

I'm not thinking about the same kind of spurious signal you are here. I'm thinking about the mass equivalent of the energy in the capacitor. This being pushed back and forth one way would produce a signal in the balance different from pushing it in a different direction.

When I first looked at this thruster idea, I too assumed that the mass being pushed was the mass equivalent of electrical charge on the capacitor. I have since been advised that this is not the case. The mass equivalent of the energy charge on the capacitors is too small to produce a significant effect. I have been told that it is the transition state which creates the mass fluctuation, not the quantity of charge present on the capacitors.


In other words, it is the act of charging or discharging that creates the change in mass, not the actual charge, or lack thereof.


This is my understanding anyway.

[GIThruster]

In other words, it is the act of charging or discharging that creates the change in mass, not the actual charge, or lack thereof.

This is my understanding anyway.

That is almost exactly right. It is the combination of a change in internal energy, with an acceleration relative to the distant stars, that causes a Mach Effect or mass fluctuation. That fluctuation occurs at the 2w of the change in energy. So, if you are driving a capacitor that has a 1w piezo response at 40 kHz, it is charging and discharging and accelerating left and right at 40 kHz. the Mach Effect is then generated at 80 kHz. In soft PZT, the electrostrictive response is 2w or 80 kHz, so that acceleration is what "rectifies" the change in mass into a net force.

PMN has no piezo response but is more highly electrostrictive, so it can't be driven with a simple sine wave. (Actually, for change in energy, a sawtooth or triangle wave will work even better, but the power system is not there yet.) So for PMN, the signal into the thruster needs to be a more complex, 1w + 2w wave, but the phase between those portions can be altered, demonstrating a new level of mastery. And the ability to switch the direction of a thruster is mighty handy besides.

Also note, it is because it is the change in energy that generates M-E, that transients such as those caused switching the system on, generate spikes in thrust. The sigantures of the devices support M-E theory and do not support the QVF model at all that I can see. Else Eagleworks would have gotten thrust from their DC studies, and their largest thrusts like the 110uN impulse, would not have been caused by transients.

[GIThruster]

double post

[Diogenes]

PMN has no piezo response but is more highly electrostrictive, so it can't be driven with a simple sine wave. (Actually, for change in energy, a sawtooth or triangle wave will work even better, but the power system is not there yet.) So for PMN, the signal into the thruster needs to be a more complex, 1w + 2w wave, but the phase between those portions can be altered, demonstrating a new level of mastery. And the ability to switch the direction of a thruster is mighty handy besides.

The piezo material is non-linear? Shades of Soliton!


So if you drive it with a sine wave, you get a non sine wave resultant? Yes, that does make it more difficult to drive it properly, but you can still use RF coupling to transfer the energy to a wave shaping circuit.



This is rather odd, because piezo crystals are used all the time to produce sinusoidal waves. I'm wondering if you can't simply configure your piezo material as some sort of power oscillator and let it find it's own resonate frequency instead of trying to drive it from a non resonant input.


I'm a big believer in feedback.

[GIThruster]

If the resonant frequency isn't known within a narrow boundary, direct RF drive is probably not feasible.

There was a scheme back around 2007, to power a microwave thruster in a kitchen oven. Apart from the inability to sweep the frequency and adapt it to the changing mechanical resonance (it changes as the device heats up, along with the capacitance and thus the electrical resonance) there is the trouble with instrumentation. How do you trust a simple accelerometer when you're pounding it with rf?

But if all you want is a cheap demonstration of the effect and no real scientific controls, such a thing is certainly possible. You can't learn much from such an experiment. You can have fun though.

I did hook Jim up with a PLL generator that I hope will work for him. He has never had PLL matching before. All the work has been with manual Z matching via a hand cranked inductor, and mostly sweeping the thruster through it's current mechanical resonance to see it spike when hit. Jim has manually tuned to the resonance and this is what produces the most spectacular results, but PLL would put him hugely ahead, and I think it was PLL that was responsible for the large signals generated at Eagleworks.

Eventually, a commercial iteration will have active thermal management so it does not change temperature or mechanical resonance, and the reaction mass will then be able to be tuned to that resonance in order to produce a much larger mechanical acceleration--basically an acoustic Bragg mirror or reflector. With the thruster operating at a stable frequency, z-matching also becomes quite easy. So we do know the way forward to solve the difficult issues, but these things are not cheap and the work is still unfunded.

[vnbt4]

My goal is to build a qvpt as named by eagleworks in hopes to get it to hop off a table. Though it seems such a result is unlikely.

[GIThruster]

I'm wondering if you can't simply configure your piezo material as some sort of power oscillator and let it find it's own resonate frequency instead of trying to drive it from a non resonant input.


I'm a big believer in feedback.

That is one of the current challenges. The thruster needs to be driven on electrical resonance, which requires a variable inductor to be tuned. Eventually we'll want to see "high speed auto z-matching" for this. It's not a difficult thing but it is not cheap either.

The thruster also needs to be driven on mechanical resonance, just as you suggest; and that is when we see the largest thrust signatures. Because there is no active thermal stabilization in the current design, the thruster changes temperature and when it does, its resonance changes. There are all sorts of negative consequences to this, but the primary two troubles here are that a) the active material does not go through its largest extensions and hence produce the largest accelerations, unless it is in mechanical resonance, and b) the reaction mass is not tunable. It reflects a single acoustic frequency and when the thruster is not on that frequency, the reaction mass does not cause the excursions to develop constructive interference.

However, there are fixes to all these things. Most just take some time and money to develop into test articles. Any BS EE could design a PLL circuit for the current power system that would remove the need to Z-match by hand, based on designs and instructions I could give them, but this takes time and the work is unfunded.

[hanelyp]

Driving a piezo element at mechanical resonance is a fairly old and very well established technique, widely used as a frequency/time base for radio and clocks. Driving a capacitor in time with a clock isn't difficult, but works better if the capacitor is in a tuned circuit close to the clock frequency. Depending on circuit Q the last few percent of frequency match may give minuscule gains to the drive. Quadrature phasing is easy with the ne567 IC, and other phase shifts can be had. Clean up that antenna of an AC feed and phase shifting on the capacitor can become a control factor.

[GIThruster]

A sintered stack is nothing like a piezo clock. It changes capacitance and resonance when run dynamically, so it is a significant challenge to keep it in resonance. The best way to meet the challenge is with PLL. Jim is working on that. I'm hoping he'll use the Labview solution. Anyone interested to do some graphical programing with Labview, using a dual VCO with saw and triangle, and a PLL, please let me know.

The changing resonance is a problem that needs to be overcome to have a new level of mastery. Currently, it is used as a control, since as one sweeps through the current resonance, one can see the thrust peek and reversal (when the 1w and 2w go in and out of phase.) Better though, is to control the resonance deliberately, which is what PLL offers.

People have been all over the lab with rf meters. The setup isn't broadcasting.

I think one thing you don't appreciate is that Jim has been doing his own circuit design, construction and programing for several decades. Others have helped these last few years. Paul March ran some new circuits on circuitmaker a couple years ago, and Bruce Long (whom I met here at T-P), the Penn State PhD EE, built Jim some stuff last year. So it's not as if the talent doesn't exist for this stuff. The trouble is more an issue of manpower. I believe Eagleworks has 4 full time EE's and has had a couple interns as well, and it still took them 2+years to accomplish what they have. Most of the time, Jim works alone. If there were funding, the first need is to get an EE in the lab so these things can be accomplished on a more timely basis.

[ladajo]

Phased Locked Loops have been around for decades. What is Jim's limiting factor in applying it to his rig? Or is it just a factor of not having done it before?
I think it is great he is going for more control, but I am unsure other than that what the magic is in using PLL to control the oscillators. I can't see it being a frequency issue given the applications of PLL in radar and sensing. Dunno, just askin. I remember playnig with them back in the day with vacuum tubes. Good fun. Much easier now with solid state.

[GIThruster]

Phased Locked Loops have been around for decades. What is Jim's limiting factor in applying it to his rig? Or is it just a factor of not having done it before?

That and the time to do it. PLL is not very complex, but when Jim is in the lab and the setup is working, of course he prefers to gather data.

He is talking about doing the PLL himself, however. I found him a dual VCO and PLL circuit that might serve, but IMHO, the best implementation of this is in Labview. I sent all that along too. There are other issues that Labview and PLL might address. When one looks at the accelerometer data, it's obvious that there is a lot of noise in the system, and the ceramic is not accelerating as uniformly as one would like. One benefit from Labview is that one might use active noise canceling to add back the inverse wave of the noise and cancel it completely, much as noise canceling headphones and bluetooth mouthpieces do. This would have lots of benefits from enhancing the s/n to keeping the ceramic cooler. This can be accomplished using the Labview Adaptive Filter Toolkit.

I'm a huge fan of Labview. Here's a cool vid of something like what I want to see in the lab:

https://www.youtube.com/watch?v=5Av4D2AB2UI

and here's how to do PLL in Labview:

http://ece.umaine.edu/ece/files/2012/06 ... report.pdf

[ladajo]

Some of the industry (R&D) PLL boxes are configured for Labview from the factory. Sounds like what you found him may be as well.

He may need more channels though. But you can find them cheap enough poking around. Or, as you said he can make his own. It is the smaple rates that drive costs, but I can't imagine he is looking for GHz sampling on the PLL.

[GIThruster]

. . .I am unsure other than that what the magic is in using PLL to control the oscillators. I can't see it being a frequency issue. . .

The biggest issue with PLL is that, as the stack heats, the mechanical resonance changes and when driven off resonance, the thruster does not develop the extensions and attendant accelerations it otherwise would. The PLL would be to adjust the frequency of the single VCO in the case of PZT, or both VCO's in the case of PMN, to the natural resonance of the clamped stack as it changes. The acceleration is used both to generate M-E (1w) and to rectify it into force (2w), so operating on resonance has a quadratic advantage. It also requires far less power and generates far less heat. For example, commercially available monolithic stacks that generally dissipate 100W, dissipate only 1w when run on resonance. So you can guess their thermal management is much better, and they can be run for much longer periods of time without active thermal systems and without depolarization.

PZT is polar and depolarizes when heated over Curie. That will fry a thruster and repeatedly has. PMN is not polar so if it overheats, no harm done. Too, PMN has a k value 20X that of PZT, and thrust scales quadratically with k. Still waiting on the new PMN discs to arrive.

He may need more channels though. But you can find them cheap enough poking around. Or, as you said he can make his own. It is the smaple rates that drive costs, but I can't imagine he is looking for GHz sampling on the PLL.

This is an amazing little box. Count me impressed at how simple NI has made this, and you're right--Jim would need a lot of channels since he records data from 3 accelerometers for each thruster, temperature, current, voltage and of course thrust. I'm sure there are others I'm just not thinking of.

https://www.youtube.com/watch?v=ofzbA3k ... 5D7BB78F9C