Parametrics, Noise coherence, and Switching

[ion]

From Orthofield:

In general, I find that free energy inventors tend to ignore small, certain gains, in preference for large, possible ones. If one of our principles is to maximize the conventional efficiency as much as possible, so too should it be to utilize any definite gains, no matter how small. Just like static electricity in the 17th century, we have no idea where this will go...

I agree we should explore and develop even small effects, provided we have not already explored and mapped out the absolute practical limits and found them to be unworthy of further pursuit at this time. They may need to await an advance in material technology.

[T-1000]

From Orthofield:

I agree we should explore and develop even small effects, provided we have not already explored and mapped out the absolute practical limits and found them to be unworthy of further pursuit at this time. They may need to await an advance in material technology.

Well, to study the topic you can start with bringing power transformer near Tesla coil's inductor or other high voltage capacitor disharge pulses source. Like in following video it shows resulting energy amplification effect and drop of power usage in primary circuit:

The second use case is when you recycle reactive power back to circuit and inject it as supporting magnetic field in resonant way:

These leads can be very exciting when you begin understand fundamental principals behind it...

Cheers!

[Orthofield]

Hi T-1000,

Interesting tests!

I've seen different versions of the kind of "reactive power" device in the second video a couple of times, and know from my own experiments that it can be done.

But the first test is a strange one. It reminds me most of the Correa's more recent work, where they suggest that 'massfree' Tesla longitudinal waves arise from the monopole HV antenna or electrode. At least some part of these waves are not electrical, and this part can modify matter and conventional energy. Here's quote from the difficult to follow US patent US7053576:

"Essentially the first subtype or variant consists of longitudinal massfree waves that deploy electric energy. They could well be called Tesla waves, since Tesla-type transformers can indeed be shown experimentally to radiate massfree electric energy, in the form of longitudinal magnetic and electric waves having properties not reduceable to photon energy or to 'electromagnetic waves', and having speeds of displacement that can be much greater than the limit c for all strictly electromagnetic interactions."

But even on 'just' an electrical level, your probe can be seen as one plate of a capacitor, with the wire as the other plate. The subject of electrical fields and charges on the surface of a conductor is not well known, but it's true that the surface charges and fields control the current in a closed circuit. See the interesting article by Jeffimenko attached. It should not then be surprising that additional fields such as provided by your probe into the transformer would change this current.
I think the conventional explanation for the null effect of an outside E field on a current assumes the added E field is a dipole, but I'm not sure that this is the case in your test, where it seems the E field from your probe (or one-wire transmission) into the coil is more of a monopole, with the other end somewhere distant from the coil-- is that right?
In the case of a monopole E field inside the coil, it could have an effect on the surface charges. Hmm, especially if the oscillating single wire E field is 'pumping' current across the whole coil in some way..

A guy named Elouard did some seemingly similar things in France in the 40s, where he fed static electricity from an antenna or Hv generator into a cone shaped assembly and used it to 'accelerate a current'. I can't retrieve that patent at the moment..

Then there is also a patent from Burke for a wire around a radioactive coil that generates has more current going through it. Burke is attached:

"this invention relates to the amplification of electric current by the emissions from radioactive material through which current is caused to flow, and also is that the electric discharge of the radioactive material is enhancedby causing an electric current to flow through the radioactive material. It is not known how this amplification physically occurs, but it is thought most probably that the radioactive material excites or stimulates the input current. In any event, the term amplification is intended to mean merely that the output current from the radioactive material is larger in terms of amperage than the input current as from source 4 in FIGURE 1."

Even though radiation is a bit different from oscillating E fields, both of them do seem they could be carriers for 'mass free fields'.

This is pretty off topic for this list, so I've started a separate thread for discussion of the anomaly shown by T-1000 in his first video. Or perhaps there is already a thread on this, because it looks like an interesting topic...

orthofield

[T-1000]

More on same thing:

It is Tesla coil primary from the left and push-pull inverter from the right in video. With capacitor discharges in the left that coil is making explosive impulses which affect transformer and interrupt "normal" sinus wave flow which is making bulb to be lit much more brightly.. If you will repeat that experiment there will be good starting point to make other attempts with same principle eventually leading to unknowns...

Cheers!

[Orthofield]

Hi ION,

>I agree we should explore and develop even small effects, provided we have not already explored and mapped out the absolute practical limits and found them to be unworthy of further pursuit at this time. They may need to await an advance in material technology.

Yes, I realized that my statement before was not a balanced one. When effects are known to be small in magnitude, it may not be worth the man hours to investigate them, relative to other areas that may be lower hanging fruit.
Shanefield itself is not as promising as other things.

So I return to the general subject at hand...

We've discussed a couple of different concepts that can convert noise or thermal energy into usable power. I see several possibilities for a powerful thermal or noise coherence device, but the one that seems most promising to me right now involves transformed feedback loops, as shown in the Black patent.

There is the possibility of a variety of regenerative or feedback effects possible, which may result in energy gain. The condition for genuine energy gain in Etotal - Efeedback supply > COP 1, and I consider this even possible for an active device.

This energy gain can exist due to the noise from an amplifier connected to a resistor being amplified and transformed and fed back to the original noise. But I also noticed that even feeding back a part of an transformer's sine AC output to a third coil that opposes the primary and secondary fluxes, improves the transformer performance. This is the principle of the Cobb Energy Conservation circuit:

http://www.affs.org/html/the_energy_trimmer.html

and the attached patent. This device was in commercial use to reduce electric bills in California. (I don't know if it still is..)
Feeding back current to a third coil opposed to the primary increases the primary input impedance, and decreases the secondary output impedance, to the advantage of the output power:

"Since the third winding produces magnetic flux in a direction opposing the magnetic flux of the primary winding, the presence of the third winding increases the impedance in the primary winding. This increased impedance results in a reduced current flow in the primary winding, thereby reducing the current drain from the public utility power supply lines. Since the current flow in the secondary winding is induced by the current flow in the primary winding, the secondary winding acts as a load with respect to the primary winding. The electrical current flow in the primary winding produces a magnetic flux in the core which causes a flow of electrical current in the secondary winding. Since the third winding also acts as a load with respect to the power coming in, the flux induced in the core by the flow of current through the third winding is in a direction opposite to that of the secondary winding. This flux produced by the third winding reduces the impedance in the secondary winding so that the electrical current in the secondary winding is not reduced by the reduction of current in the primary winding that results from the presence of the third winding.
The presence of the third winding on the load isolation transformer produces a magnetic field which decreases the energy input from the primary winding, but keeps the energy output in the secondary winding the same."

Feedback in this sense has possibilities that go well beyond what can come from noise power, even at high BW. The odd part about the above is, taking the patent logic to its conclusion, ALL the primary and secondary fluxes could be cancelled in a transformer, without any expense in the output! In this case, the primary input impedance is infinite, and the secondary output current is the same, since the output impedance is zero! Of course this cannot happen in reality without feeding some energy in, making the system not passive. But how close can this be approached by using feedback windings, perhaps transformed in the ways I see used in the Black patent?

Consider that coil 26 in the Cobb patent is usually directly coupled on the pot core to oppose the secondary and primary fluxes. But I can imagine improving the output characteristics even more, by magnetically uncoupling coil 26 from the pot core, using a new winding with a few turns around 26 and a few turns coupled to the pot core and primary and secondary. This new coil steps down the voltage from 26, getting a higher current which can then be set by turns ratios to to match the primary current. The reverse flux of the third coil can completely cancel the first coil's flux.

Is this possible?? Or have I finally lost my marbles, as many claim? ;-)

In any case, the important thing to note in my wanderings is that the Cobb device is a passive feedback loop. It doesn't require an amplifier in the middle of the transformer chain as does the Black Patent. This implies that there are unexpected possibilities even in passive feedback loops, without noise.  

orthofield

[Orthofield]

Hi All,

I'm restarting this thread after re-evaluating the attached patent. I think it's relevant to the topic of parametric amplification through switching of an inductance. Those who've read earlier posts on this thread know that Jean-Louis Naudin did an experiment I suggested and got a mV sine wave result from two switched inductors, with no power input. Later the test was repeated with an opto-isolated switch and got similar results with high harmonics indicating parametric processes. Further, the Barrow research from the 1930s referenced earlier in this thread indicates that switching of a capacitor can induce oscillations in a tank circuit when using a regenerative circuit to maintain low dissipation. 

I still consider this a fruitful path for cohering thermal noise vibrations into useful power.

I've had this patent in my files for a long time but somehow the significance never came to me.

The Variable Reactance Element was patented in 1968 by Lucio Vallese and assigned to ITT. The device is intended to be used in a Parametron, or parametric oscillator that was in development at the time for computer logic and RAM. After pointing out the drawbacks of varying reactance with vacuum tubes and magnetic materials with nonlinear BH curves (what we would call mag amps or parametric transformers), he goes to the operating principle, described on pg. 3, col. 2, line 10:

"Therefore if a value of the variable resistance element 5 is chosen such that the total secondary loop resistance
is equal to the self-reactance of inductance element 4, small variations of the resistance element 5 from this value
will result in corresponding variation of the reflected capacitive reactance X, while having substantially no effect on the reflected resistance R'. It may be readily shown that the capacitive reactance X is always less than the self-reactance of inductance element 3, so that the reactive component of the driving point impedance seen looking into terminals 1 and 2 is always inductive."

He goes on to say that this basic circuit has low Q due to the reflected resistance at the primary, and provides several solutions using a negative resistance in first the secondary, and then in both secondary and primary circuits.

From a computer logic standpoint this makes sense, but if we are looking for amplification it's not ideal. Fortunately he provides a final solution which is perfect for our experimentation. On pg. 4, col. 3, line 30, he says:

"FIG. 3c shows an alternative technique for compensating the reflected resistance R. In this approach, each of the inductance elements 3 and 4 is divided into two operating portions 8 and 9 and 10 and 11 respectively. The sign of the mutual inductance between portions 8 and 10 is opposite to that of the mutual inductance between portions 9 and 11. The circuit parameters are selected so that the real components of the reflected impedances cancel while the imaginary components add."

So to sum up, Vallese provides a method for varying the L of a parametric oscillator using only switching power, where the resistance is not reflected into the oscillator. The Q of the oscillator is then mostly dependent on whatever load in included in it.

Since power in such oscillators is dependent on the frequency of oscillation, it seems likely that at some rate of inductive variation, the power in the oscillator will be greater than the power needed to drive the switch.

(The gain is limited by the Manley-Rowe relations but this can also be circumvented by including another switch in the circuit run at quadrature to the switch in the secondary loop, as described in papers I can make available to anyone who wants to try this out.)

Fred 

[Orthofield]

Hi All,

More evidence that L or C in an oscillator can be changed by only switching or varying resistance, without energy being lost in separation of flux or charge, and without the varying resistance reflecting into the oscillator.

The thesis "The Use Of Transmission Lines As Frequency Modulators" by Earl Dennis Scott, University of Washington, 1933 is about this exact topic. The author proves mathematically that this is possible, in a similar fashion but at much greater length than in the patent by Vallese, and then tests the concept with an artificial 1/8 transmission line consisting of wire wound around a dowel covered with aluminum foil and waxed paper.

His claim is:

"It therefore seems possible that a line might be chosen of such a length that a change in resistance alone of the terminal impedance would result in a change of reactance alone at the sending-end impedance. If the sending end of such a line were then associated with the tank circuit of an oscillator, and the terminal resistance were varied, frequency modulation of the oscillatory current would result without amplitude modulation."

You can see that this is precisely what the Vallese patent does through other means.

As with the Vallese patent, he shows test results. It's difficult to see in the attached image, because of a defective Xerox machine, but as the R at the receiving end of the artificial line is varied from 260 to 340 ohms, the other end varies from an Xc of about 30 ohms to an Xl of about 60 ohms.

Although he uses this for frequency modulation of a Hartley oscillator, it could just as readily be applied to creating oscillations in a tank circuit without any other input. Using only the L or C range of the T line, one creates an oscillator with fixed C or L, and some small R, and commence to rapidly vary the R at the other end of the line, at twice the tank frequency. A basic parametric oscillator. As with the Vallese patent, the varying R is not reflected into the tank circuit, so only whatever load is present reduces the Q of the oscillator.

I have other documentation that more obscurely shows the same results. I take it is as a given now that parametric oscillations can supply a load with power, with the input power only that needed to vary a resistance (ideally between closed and open circuit).

It may be objected that it takes a minimal energy to switch, and that the output power of such an oscillator is limited by this minimum. This is called Landauer's principle, and is directly related to the second law of thermodynamics, and to the existence or nonexistence of a Maxwell's Demon.

It has been demonstrated experimentally that Landauer's limit can be violated by using processes similar to those in adiabatic computing. The attached article shows that one can indefinitely decrease the energy consumed by moving charge to and from a capacitor well below the thermal dissipation predicted by Landauer. Adiabatic computing circuits do this as a matter of course. As the synopsis of the article concludes, "This shows that there is no fundamental lower limit to energy dissipation in moving charge."

It is only for someone with more building skills than myself to build such a circuit and observe the results. The results, I predict, will be an operating Maxwell's Demon of high efficiency that will cohere noise power into useful power, with limited energy input.

Fred

[F6FLT]

...
So to sum up, Vallese provides a method for varying the L of a parametric oscillator using only switching power, where the resistance is not reflected into the oscillator. The Q of the oscillator is then mostly dependent on whatever load in included in it.

Since power in such oscillators is dependent on the frequency of oscillation, it seems likely that at some rate of inductive variation, the power in the oscillator will be greater than the power needed to drive the switch.
...

Fred

Hi Fred

I don't think so. The author uses the classical equations of a circuit to obtain his equation (1), .i.e as if the resistance were constant.
If R' is not constant but varies only slowly with respect to one period of the signal, its equation is still valid.
But if we hope to make R' vary at a rate of the same order as the frequency of the signal, equation (1) becomes false, we can no longer speak of a resistance R' nor of a capacitance C', we have to go through the integration over a period of the signal, of the instantaneous relations between V(t), I(t) and Z(t).
So his idea is good for the use he wants to make of it, for example to modulate a reactance as we can do it to make frequency modulation because the modulation signal is slow compared to the carrier signal. On the other hand, all the calculations have to be redone if we intend to use the circuit for parametric amplification, and then I don't think we'll find any energetic interest in it.

Another point, I agree with the Applied Physics Letters paper : we can charge a capacitor with less losses than Kb.T.ln(2) (but as the authors have shown, at a slow loading speed).

[Orthofield]

Hi F6FLT,

Thanks for reading my long posts.

Certainly you could be right about the Vallese patent. Several factors work against a purely mathematical dismissal. It would have to be tried to be certain.

1) The patent is intended for uses in a parametron. Why would ITT have bothered to patent it if it didn't work in parametrons? I have another patent that does more or less the same thing, that's assigned to the US government. These are not gullible investors paying for a golden unicorn.

2) In the first few paragraphs the inventor specifically mentions that reactance tubes, magnetic cores, and electromechanical reactors are only suitable for low frequencies, and his device is a replacement for them. The other patent, yet to be discussed, makes the same claims.

3) parametric oscillations-- and gain-- are easily achievable at frequencies of a couple of Mhz, or even below-- the same frequencies as frequency modulators. And it's possible to get gain in even a one shot application, at quasi-DC. It simply won't build like an oscillation would. For instance, resonant transfer circuits where energy is shuttled between two capacitors through a variable inductor, or vice versa, are commonplace in the literature of this time period (50s-70s approx.)

4) This isn't a case of having a theory and looking for evidence of it. Barrow's experiments at MIT in the 30s were my original inspiration. The paper is earlier in this thread. In those experiments he saw parametric oscillations in circuits using a capacitor shunted in and out of the oscillator with a rotary switch. These experiments were done at a few kHz at most. They were critiqued in a footnote in a paper by Mandelshtam and Papaleksi as violating the law of conservation of energy, so of course I went looking for them :-) They used a regenerative circuit to maintain a dissipationless state, so they were not complete proof, but other details indicated that parametric oscillations were forming under far from idea conditions-- for instance where half the charge stored in the capacitor was being lost with each switching operation. My own experiment with Jean-Louis Naudin was crude but showed a clear sine signal in the mV range at the switching frequency, with no source of power. It was rightfully critiqued for using a relay which could have injected energy into the circuit. so it was later repeated with an opto-isolated switch, once again showing high harmonics typical of parametric oscillations. I recently found a few more details on this test and was planning on posting it here.

5) My experience in several experiments with various parametric devices is that oscillations WANT to form. Although controlling them or getting the frequencies you want is not easy, the natural phenomenon of parametric oscillations arises readily in many situations. So looking for difficulties where they have not already presented themselves is going against the flow of the observational record. It's saying, no, parametric oscillations won't form here, although they seem to occur in many situations, wanted or unwanted.

Of course, you may be right, and only more experiments will show for sure. Unfortunately, I am a dyslexic experimenter with limited equipment. I'm still working on getting some simple steps in my thermoelectric test correctly. As much as I can understand a paper or patent, I get the black wire mixed with the red wire on a pretty consistent basis. I'm really bad at the test side of things! I could build the device but would not be able to measure power output accurately. So my posts here, like really all my posts, are intended to 'sell' the idea and get people interested in the possibilities. In other words, I'm just happy if someone reads it and responds, and you are the first to do so.

In terms of the Boechler paper, you're right that it's about the speed of the charge transfer. You can charge a capacitor with zero thermal losses (including the wire or other resistance in the circuit) if you take an infinite time to do it. Losses can be reduced appreciably even if the process is just done more slowly than usual. The interesting thing to me is that 'slow' is only relative to the internal clock of the system, in this case the RC time constant. People get stuck on the word 'slow' as if it were absolutely slow, but it's only relatively slow. In my opinion, many technologies can be made far more efficient than they are by using these principles, which so far have only been used in adiabatic computers, and a few supercapacitor charging circuits. There's no reason they couldn't be used in long distance power transmission, for instance. But that's another topic, which I've covered in a thread in my bench here...

Fred

[Orthofield]

Hi All,

Continuing the discussion about energy gained through switching, I have fragments of a bit of research conducted some 22 years ago now. Unfortunately large chunks of it were on computers that bit the dust years ago and I didn't think to recover them. I apologize in advance for the missing details.

But basically, first Jean-Louis Naudin and I did the test described here:

http://jnaudin.free.fr/html/tep62par.htm

As you can see there is no power output in the tank circuit, consisting of L1, L2, and C1. L1 and L2 were the two halves of a transformer.
L2 is switched in and out of the circuit with a reed relay. The observed output across the tank is around 2 mV p-p and in phase and same frequency as the switching impulse.

Some said that the sine wave arose from some energy coming into the circuit from the reed switch. So I looked for an opportunity to repeat the test with a better switch. Eventually, the test was repeated with some modifications.

Rather than inductors, capacitors were used.  Several were tried with different ratios of C1 and C2, with varying results. Unfortunately the values and types of capacitors that got the best results are lost in time.

An opto-isolated mosfet relay was used in place of the reed relay.

I don't remember or have record of the fixed inductor used, but the tank frequency without C2 was 50.3 Khz.

The relay was driven by a 555 circuit at 100.6 Khz with a 50% duty cycle.

The output were a series of small ringing pulses. Each individual pulse had a frequency of around 1 Mhz-- but the odd thing was the pulse frequency was 8.3 Khz--which is the 6th subharmonic of the tank frequency!
It occurs to me now that the peculiar frequency may be related to the value of one capacitor without the other, but can't check that now.

Measurements were made across a 100 ohm resistor.

It's yard to tell the size of the sine wave from my very poor images (not worth reproducing here) but assuming the gradation I see is 1 mV, then once again these waves were about 2 mV p-t-p.

Without clear information there's no point in measuring the power dissipated in the resistor, but it was more than nothing. Perhaps peak .01 uW if the p-p amplitude is treated as 1 mV.

In an attempt to find error, we sent the 555 output directly into the resistor through the relay without the tank, and this only reproduced the output signal with a bit of additional noise.

The presence of a ringing pulse in the circuit at a frequency completely distinct from the tank or switching frequency is a strong indicator of parametric oscillation in my book.

Once again, I apologize for the missing details. 

Fred

[chief kolbacict]

http://jnaudin.free.fr/html/tep62par.htm
I did something similar about ten years ago.
Only the circuit was more complicated, and the inductance was with one winding, but there were two switched capacitors. And during one inclusion something happened, the circuit went into spacing and burned out.
It was the only time in my life when something incomprehensible happened.
And that I can not give an explanation until now.
p.s.
This layout was partially restored, but not destroyed or taken apart so that I would not be accused of lying.
But there were no more experiments.

[F6FLT]

TR1 is not powered, where does the energy seen on the scope come from?
It's quite obvious: from the capacitive coupling between the coil of the X1 reed relay and the contacts it controls.

Switching a non-powered inductor has no effect. For example, one might think that short-circuiting a few turns of a coil through which the magnetic flux of a permanent magnet passes amounts to varying the quantity of flux to which the coil is sensitive, which would generate an induced current. This is not the case. Connecting together points of  conductors that are already at the same potential, or doing nothing, is the same thing.

[chief kolbacict]

In any circuit, any circuit, there are current fluctuations, for example, thermal noise.
We can amplify them parametrically.

[Orthofield]

Hi F6FLT,

Yes, this was the explanation given by several at the time of this test, and it was certainly a possibility.
That's why the test was later repeated with an opto-isolated Mosfet switch, and this test also gave rise to anomalous pulses/oscillations. Putting the voltage from the CMOS directly into the tank circuit didn't result in oscillations.

These pulses were at a subharmonic of the tank frequency, and with a 100 ohm resistor, so had to arise despite considerable impedance.

This is the paper by W. L. Barrow that led to my interest in the first place.

https://www.overunityresearch.com/index.php?topic=2265.msg45496#msg45496

I'd be interested in your comments on this paper.
The test setup is on pg. 4. Note that the rotating element is a switch, not a rotating capacitor, as one would expect in a rotating parametric generator.
Pg. 8 shows current in the test circuit at different frequencies of capacitor switching. Note that the highest current is at approximately twice the frequency of switching.
Fig 1 on pg. 3 and fig. 9 on pg. 10 can be compared to validate that the device shows oscillations where parametric theory would suggest they should be found (stable and unstable solutions to Manley-Rowe relations).

There's no question that parametric oscillations arose in this circuit when the capacitance was merely switched, rather than varied in value. The sticking point is that a regenerative circuit was included to eliminate dissipation. It can be argued that this regen circuit supplied power to the circuit above and beyond what was needed to cancel resistance. However, the circuit also has two sources of loss or attenuation that could be eliminated or reduced under other conditions. First, as in all parametric circuits, any nonlinearity in circuit components will result in the limitation of the amplitude of oscillations in 'unstable' regions, which would otherwise grow without limit, and the regen circuit has these limitations. Second, at certain frequencies, the capacitors have opposite charges when they reconnected and there is a loss of energy in those regimes. Eliminating some nonlinearities in the circuit, and using a dual push-pull topology so that charge is shuttled to another similar circuit rather than being dissipated in opposition would improve the performance.

Something like this test has now been done three times with different setups, and each time there have been oscillations. So there's no reason to dismiss this line of research, and reason to pursue it.

Fred

[F6FLT]

In any circuit, any circuit, there are current fluctuations, for example, thermal noise.
We can amplify them parametrically.

The noise is several orders of magnitude smaller than the signal we see, and it is random so that it cannot be parametrically amplified with a signal of constant frequency.

[Orthofield]

http://jnaudin.free.fr/html/tep62par.htm
I did something similar about ten years ago.
Only the circuit was more complicated, and the inductance was with one winding, but there were two switched capacitors. And during one inclusion something happened, the circuit went into spacing and burned out.
It was the only time in my life when something incomprehensible happened.
And that I can not give an explanation until now.
p.s.
This layout was partially restored, but not destroyed or taken apart so that I would not be accused of lying.
But there were no more experiments.

Hi Chief,

That's very interesting! And was this circuit unpowered, as in the Jean-Louis test?

Fred

[Orthofield]

F6FLT,

I forgot to mention a key point in the Barrow paper. Fig. 6 on pg. 8 shows three currents. Ic is the current between the fixed and switched capacitors. The tank current is It. Ip is the vacuum tube plate current. Ignoring the capacitor current, there is still a difference between the current supplied by the tube and that existing in the circuit as a whole, implying another current source than the regen circuit. At peak resonance, twice the switching frequency, the plate current is around 15 mA, and the tank current is a bit more than 40 mA.

Fred

[chief kolbacict]

That's very interesting! And was this circuit unpowered, as in the Jean-Louis test?

Fred

Hi.
My old circuit consisting of an inductor and one of the two capacitors swayed slightly at its resonant frequency.
The frequency was chosen around 30 hertz, so that the reed relays had time to work.
The relay switched the second capacitor, synchronously with this frequency.
At one moment there was an anomalous increase in the amplitude of oscillations of the power circuit contur.
This led to the burnout of the CMOS logic circuits, which were separated from the power part by the relay.
I still don't understand how this could happen.

[F6FLT]

...
There's no question that parametric oscillations arose in this circuit when the capacitance was merely switched, rather than varied in value. The sticking point is that a regenerative circuit was included to eliminate dissipation. It can be argued that this regen circuit supplied power to the circuit above and beyond what was needed to cancel resistance. However, the circuit also has two sources of loss or attenuation that could be eliminated or reduced under other conditions. First, as in all parametric circuits, any nonlinearity in circuit components will result in the limitation of the amplitude of oscillations in 'unstable' regions, which would otherwise grow without limit, and the regen circuit has these limitations. Second, at certain frequencies, the capacitors have opposite charges when they reconnected and there is a loss of energy in those regimes. Eliminating some nonlinearities in the circuit, and using a dual push-pull topology so that charge is shuttled to another similar circuit rather than being dissipated in opposition would improve the performance.

Something like this test has now been done three times with different setups, and each time there have been oscillations. So there's no reason to dismiss this line of research, and reason to pursue it.

Fred

Hi Fred
I agree with your analysis. The author uses a UX-112A triode. I found the datasheet here on page 25 of the text, 29 of the pdf.
https://www.nmr.mgh.harvard.edu/~reese/RC10/rc10.pdf

The gain is 8.5, which is low compared to modern tubes but can go well beyond loss compensation. Inter-electrode capacitance can be neglected at the frequencies it uses.
The circuit can very well be seen as an oscillator with gate/anode feedback. If the feedback level is insufficient for oscillation, then there may be only loss compensation, but this is impossible to ascertain. One could also have oscillation with C0+ΔC and not with C0 only, or vice versa, I hope the author has checked all this, I don't have the impression that he has specified it.
Also the tube characteristics are not very linear. We would have to put all this into LTspice and see what happens, but just creating the triode as a component for LTspice is tedious.
I don't think I can be very helpful because there are too many uncontrolled parameters because of the tube, and we don't know the currents in the two halves of L. I don't see why the tube should be considered a simple loss compensation device when its amplification beyond that is the most likely cause of oscillations.

[Orthofield]

Hi F6FLT,

Nice find! I never thought to check for the characteristics of the tube.

Since this paper and the one before it were specifically a theory and test of the 'dissipationless' situation, he certainly was aware of the deviations from the perfect state. Barrow does discuss the 'dissipationless' circuit on the same page as the schematic, and the next. He acknowledges that the circuit only provides an average value of zero resistance through a cycle, and suggests that is the reason why oscillations were not seen for higher ratios of capacitance variation, where they should otherwise have been more likely.

Interestingly, Barrow seemed to think that switching a capacitor was the same as rotating a plate capacitor, which is certainly not the case. Later papers used a rotating variable inductance, rather than a switching scheme, so he must have been made aware of his 'error'.

As a temporary conclusion to this line of discussion, here's the original comment by Mandelshtam and Papaleksi on pg. 40 of their paper "Report On Recent Research On Nonlinear Oscillations" that led me to make considerable effort to track down the Barrow paper:

"W. L. Barrow wrongly assumes that his experiments shows the capability for the parametric excitation of an oscillating circuit by periodic variation of its capacitance. He causes the variation of not only the capacitance of an oscillating circuit, but the ohmic resistance of a shunt containing a condenser. Now, the variation of a positive resistance can be carried out (and is carried out) without expenditure of energy. This device therefore does not allow, by its principle alone, supplying to the circuit by mechanical work, of the energy necessary for the excitation and maintenance of oscillations. There is no doubt that Barrow has observed in his experiments not only the parametric excitation of oscillations by periodic variation of the capacitance, but also phenomena owing to the presence of an electronic tube and regenerative feedback." 

My comment on their comment is that this is true in Barrow's circuit, but Vallese's patent and the Scott transmission line thesis, as well as another patent I haven't presented, show quite clearly that under certain conditions the ohmic resistance can be fully reflected into a pure reactance, so that variation of the resistance alone can cause oscillations in the reactive side of the circuit.
Even if this is done in a quasi-static fashion, there is still an energy gain from the process of changing the parameter.

That's where it ends for now, until I can get an oscilloscope and do some tests on (probably) Vallese..

Fred

[F6FLT]

I agree! 

[chief kolbacict]

Here is that "хрень"
It was broken nothing.Might continue.

[Orthofield]

Hi Chief,

Just some basic details would help.. what were you trying to do?

Fred

[chief kolbacict]

In the oscillatory circuit, weak oscillations were maintained at its resonant frequency.
No more than 1 volt. By supplying this frequency from this board through a resistor with a large resistance.
And the switching of capacitors was carried out synchronously with this frequency using a relay.
Moreover, the DIP switches, which are visible on the board, changed regulated the duration and delay relative to the
phase of already existing oscillations.
And at one moment, the amplitude grew abnormally and smoke had appear...

[F6FLT]

@Chief
Whenever I feed anything with a fixed frequency AC current, even a relay, any circuit tuned to that frequency and placed nearby will show a signal of the same frequency or ringing harmonics. Capacitive and magnetic coupling are inevitable and without extreme shielding precautions, no conclusions can be drawn.

Could you provide a detailed diagram of the device you experimented with and the data?