LTspice models for bifilar coils

[F6FLT]

Hi All

Here are two ltspice models of bifilar coils, as realistic as possible.
I created them to answer questions about amazing experimental results. But after analysis and discussions, it turns out that they are perfectly compatible with everything we know, and therefore modelizable.

If you do not have a magnetic coupling to your bifilar coil by a third coil, then the first "bifilarCoil" model is sufficient.
Otherwise you need the second one, "bifilarCoupledCoil". It was made to be able to couple the bifilar coil to a third one, because the ltspice models are not adapted to the use of a mutual inductance coefficient except for pure inductances, and the "bifilarCoil" model is not defined as such. The third inductance of the second model can be used indifferently to extract or inject a signal into the bifilar coil.

How to get these components in LTspice?
As I was unable to upload the files altough they are text files, because their extensions are not allowed, I put them here:
http://exvacuo.free.fr/div/Outils/LTspiceModels/BifilarCoil/ (right click the file names to download)

Place the two "asy" files in the "lib/sym" folder of LTspice (under Windows, mine is in user/Documents/LTspiceXVII/lib/sym). The asy file defines the symbol of the component, its parameters and ports.
Place the two "sub" files in the "lib/sub" folder of LTspice (in Windows, mine is in user/Documents/LTspiceXVII/lib/sub). The sub file defines the architecture and internal functioning.
These files are editable text files.

After restarting LTspice, when you create a schematic, you can access these two components like any other, just click the "component" icon, select "bifilarCoil" or "bifilarCoupledCoil" which appear in the list, and place the component on your schematic.

My next post explains how to set up the components.

[F6FLT]

Parameter settings of the "bifilarCoil" component

After placing the component on your schematic, right-clicking on the component opens the configuration window.
In the "SpiceLine" line, you will find the following default values:
L=1µH R=1 C=1nF K=0.9 Cp=10pF

The first three values must be adjusted according to your bifilar coil:
L is the inductance of each winding.
R is the resistance of each winding.
C is the capacity between the windings.

The following two default values will generally be appropriate:
K=0.9 is the mutual induction factor between the two coils
Cp=10pF is the "parallel capacitance" of the inductors. This is a parasitic capacitance identical to the parallel capacitance that can be specified in the LTspice base model of an inductance, but here the two inductances are treated together.

Warning: this Cp capacity must be specified if its impedance is not negligible compared to those externally connected to the ports. For example, if the generator is connected to the bifilar coil by a capacity of 15pF, then the 10pF of Cp are not negligible. If that's by a capacity of 10 nF, Cp is completely negligible.
Cp can make the coil resonate at 100 times the normal frequency related to C. Its value is critical if your coil is only weakly coupled to the rest of the setup.

Parameter settings of the "bifilarCoupledCoil" component

After placing the component on your schematic, right-clicking on the component opens the configuration window.

In the "SpiceLine" line, you will find the exact same line as in the" bifilarCoil" model. The parameters have the same role, see above how to fill the values.

The next line "SpiceLine2" concerns only the third coil :
Lc=1µH Rc=1 Kc=0.9 Cpc=1pF

The values must be adjusted according to your own third coil Lc coupled to your bifilar coil:
Lc is its inductance
Rc is its resistance

The next default values will generally be appropriate:
Kc=0.9 is the mutual induction factor between Lc and each winding of the bifilar coil.
Cpc=1pF is its parallel parasitic capacity

Note that you can always use the "bifilarCoupledCoil" component instead of the "bifilarCoil" component, provided that you don't use Lc (just set K=0 and put a ground connection to a end of Lc). The functioning of the bifilar coil is strictly identical in both cases.

My following post gives examples of elementary schematics using these bifilar components.

[F6FLT]

Example of use of the bifilarCoil models.
The parameters are those of my real bifilar coil and the slopes fit really well what I got in my experiments.

bifilSerie.jpg shows a connection of a generator between one end of one winding and the other end of the other winding of the bifilar coil.
The current follows the first winding, and all along it passes through the dielectric to the second winding and back through the ground.
The resonance frequency is low, entirely determined by LC. The Cp capacity does not matter.

bifilTesla.jpg shows the connection à la Tesla of the bifilar coil: the two coils are connected in series, the dielectric is in parallel.

bifilFree.jpg shows the resonance of the two-wire coil when it is very weakly coupled to the generator. The third coil is used here to inject the signal. The mutual inductance factor Kc is only 1%, so the coupling is weak. The capacitance Cp is the main capacitance that determines the resonance frequency, much higher than that of the other cases. C plays no role here, the two windings of the bifilar coil act as a single one tuned with Cp.

Note to the electronics experts:
The Cp capacity is visible on the components symbol, and it is connected to a port that you can access. This port is the common point of the four parallel capacities that modelize the parasitic capacity of bifilar coil.
Generally you can leave it free. Nevertheless as the coil is generally coupled to the environmental ground, a good idea is to ground connect this port. It is even necessary to do it in case the bifilar coil is weakly coupled and the Cp capacity plays an important role, as in bifilarFree.jpg.
If the bifilar coil would work in a specific environment, for example above a metallic plate that is "hot" question of potentials, this port can be connected to this hot point.

Note to the skeptics:
We know that no additional energy can appear from a model (except in case of bugs, mathematical limits of the processor or bad parameters). So, is it a good idea to search for free energy using softwares? The answer is "yes", for the following reason.
When you have experimental results that seem surprising to you, and that's what happened to me, you can invoke new phenomena, new physics, or you can model your set-up and test if the conventional results are in line with your experiment. If so, why invoke miracles?! As physicists say: "shut up and do the math" but here it's not even necessay, we have a slave to do the job, LTspice.   

[ayeaye]

I'm not against modeling at all. Modeling also enables to see how the real experiments differ from models, and this may tell a lot about what is going on, and maybe even enable to calculate things previously unknown.

I'm against replacing experiments with modeling though. Say one assumes that all bifilar coils work exactly like the LTspice model. This has any value only when one confirms it with experiments. All bifilar pancake coils work like the model, one thing less to research, then what about bifilar rodin coils, abha coils, etc. That they all always work exactly as the model, that will be great thing to find out. But this can only be done by experimenting. These things on the computer screen may look great, but the things in the computer are not by themselves reality, unless it is confirmed experimentally that they are, and always are.

What concerns modeling, LTspice may not be the best way to model coils. Though it can be very useful. The best for research i think would be to model coils by separate code, i know it's not simple, but it's doable. Especially when there really appears to be something that differs from typical models.

So on the bifilar coil made by the Tesla patent, the resonance is determined by the inductance and capacitance? In your case 30 kHz something. This is how i thought yes, as in that coil the difference of potential is v/2 and the capacitance matters the most.

[F6FLT]

In fact we didn't need to experiment the Tesla coil before designing the model. His patent is enough. We just need to take "real" L (with series resistance and parallel capacities), and to treat two of them by just adding the in-between capacity.

LTspice is by far sufficient to modelize a bifilar coil in an electronic context. We don't even need to use all the possibilities that it allows (for example, we can neglect parallel resistance).
LTspice is not sufficient to modelize in the context of physics. The topology of the electric and magnetic fields remains unknown. Other tools are needed for this.

No experiment is necessary to test everything we can imagine. For example, "What if I had a third winding?"  "What if I replace Cu with Al for the wire?" "What if my pancakecoil isn't flat but curved like a horse saddle?"... it would be futile. There are millions of configurations of anything without any sign that they would bring something new. We only need to test ideas for which there are strong indications of a possible new underlying principle, theoretical or practical.

For example, I started with these coils because I have the idea of producing induction currents in insulators. Displacement currents in vacuum or insulators produce magnetic fields so conversely a variable magnetic field must create a displacement current in a dielectric. The advantage is that there are fewer losses in insulators than in conductors (except superconductors) so we could realize high Q resonant systems. When the (not electrolytic) capacitors of the industry are almost perfect (low electrical resistance, almost no dielectric losses), the inductors, on the other hand, are very bad: rapid current dissipation in their resistance while a capacitor maintains its charge for a long time.
The configuration in "bifilSeries.jpg" suggested to me that part of the current that passes from one winding to the other through the dielectric could produce a displacement current partly along the dielectric. In fact, probably not at all, the electric field remains well perpendicular to both windings.
Now I'm thinking of other ways to "coil" a dielectric, to make a kind of longitudinal capacitor whose dielectric thickness would be much greater than conductor length, capable of being coupled like a coil to a magnetic field that it would encircle.

If there are thousands of feasible but useless experiments because nothing predicts the slightest interest, and in this case modelling is a waste of time but less than real experiments, on the other hand I have one that could be interesting but I don't know how to build it...