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.
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.
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. :)
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.
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... :(
Quote
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... :(
Hi F6FLT,
I wonder whether you mean how to build it in LTspice or in practice? 8)
Anyway, there is an active and friendly yahoo mail group devoted to LTspice, see here:
https://groups.yahoo.com/neo/groups/LTspice/info (https://groups.yahoo.com/neo/groups/LTspice/info)
Gyula
Hi Gyula,
I mean, build it in practice.
With LTspice, the "dielectric circuit" would be modelized by a looped LC circuit with zero resistance for L. The current being the same along the loop, there is no difference if it passes through a conductor or a dielectric.
The aim is to increase the length of the path in the dielectric and reduce the path in the conductor to near zero so that losses are reduced to near zero also. It is not a theoretical problem but an implementation problem.
Then I think it is an intended application which would mainly define implementation.
Probably the use of the dielectric material sandwiched between short copper strips could be a basic building block?
The goal is to induced a current in a looped dielectric from an inductor. A coil inductor generates a certain voltage/turn. For instance if we use a coil made of only one wire turn powered by 1 volt, we can get no more than 1V in a coupled loop of one turn, either conductor or dielectric.
If we sandwiched N blocks of plates/dielectric to make the induced loop, we divide by N the voltage between plates of each piece. It's like we had a capacitor with a stratified dielectric including intermediate plates of aluminium foil: it doesn't change anything because the intermediate plates occupy the same equipotential surfaces that would be occupied by the dielectric if the foil was absent. The potentials remain the same.
I think to make each stage resonant, but with no magnetic coupling and with in phase currents. Not yet found a solution.
F.Y.I.
F6FLT - excellent work and IMHO a very good "next level" approach to understanding
and ultimately solving many of these "Advanced Energy (AE)" questions.
Computer Aided Engineering (CAE), and in particular some of the newer more advanced
versions, can all but replace "early version" physical prototypes thus providing great savings
in wasted time and cost.
A "proof of concept" is [easily and quickly ?] created, analyzed, modified and tested long
before the "build cycle." But; it is not "paint by number" or a "connect the dots" exercise;
it's more complex up front! You have to understand what you want to accomplish and how
you plan to do it!
This is where the "rub" is created between a design engineer and an experimental physicist.
They are both researchers in the AE field with the only difference being their approach to
problem solving. The design engineer - these results sure do not match the expected
outcome, but why; and the experimental physicist - eureka, but why!
That's where another major advantage of integrating CAE into the process surfaces,
finding the "why" may be much closer at hand. It seems to work - but why - how can
it be explained (postulated) in an understandable way?
Over the hundred or more years of AE discovery and achievement this has been the
missing component - lots of "paint by number - connect the dots" and lots of "loose math
with hand waving somewhat probable ideas" but little or no credible, provable, repeatable
[believable] postulations (practical plus theoretical).
An Advanced Integrated Approach to AE Development [an idea - one of many]:
Integrate CAE into the research/build process early - as F6FLT is doing! Model the idea as
much as possible in the initial stage; advance the modelling in parallel along with the
ongoing build/measure/test phases - tie practical with theoretical through CAE.
Also, consider advanced CAE Multiphysics (recent developments in the craft are astounding).
One example worthy of study is COMSOL Multiphysics (e.g. draw a multi-layer coil in CAD and
completely analyze this coil - EM fields, resonance, etc., - subject the coil to pulsed HV, etc.).
Also, LTSpice models can be input and used in Comsol so you can leverage previous work.
Comsol allows you access to the formulas used in the analysis - you can edit these which is an
extremely huge deal; especially when considering any newly discovered anomalies!
Cost/benefit is pretty good (lots of demo and student deals) and computer requirements are
somewhat viable (or use networked cloud arrays for really complex designs). Overview:
https://www.comsol.com/video/first-look-comsol-multiphysics (https://www.comsol.com/video/first-look-comsol-multiphysics)
Disclosure: I have nothing what so ever to do with Comsol or anything related - I'm a user only.
Have a great day!
FIN
Quote from: SolarLab on October 29, 2018, 11:15:29 AM
Computer Aided Engineering (CAE), and in particular some of the newer more advanced versions, can all but replace "early version" physical prototypes thus providing great savings in wasted time and cost.
SolarLab, i'm a programmer, i'm all for using software, apparently more than others here. But i'm not for replacing physical research with software. I'm for using software to aid physical research.
Ayeaye, FWIW - seen some of your work elsewhere - impressive!
Agree that only CAE, without physical builds and verification, would simply keep us at the same point we are
now at. Lots of "speculation" and "cartoons" but no real viable progress, although, I feel an "only CAE"
approach would advance the art a little (maybe a lot) further.
An advanced CAE model, with the math exposed as can be done using a full Comsol analysis, even without
fabricating a physical prototype, would be very instructive (as compared to only a youtube video with a few
lines of explanation and some answered comments). Forum discussions are a bit better but an inter-active
CAE analysis is worth thousands of words and pictures. Modifications are much easier as well.
The combination of a built prototype, measurements and a full CAE analysis I believe is the ultimate approach
and probably the only one that will yield any useful results. I've seen a few builds (prototypes) that actually
work but I'm still non-the-wiser as to how or why they function. Without that piece of the puzzle it's extremely
difficult to replicate, improve on them or advance the concept.
But I will say the ongoing CAE analysis of a so called "grenade coil" appears to be revealing many things that are
otherwise not apparent, or discoverable, with only the physical model to work with.
Another bonus surfaces when your CAD/CAM/CAE design uses CNC fabrication - the G-Code is provided [coil winding,
PCB, fixtures, etc.]. Plus when you need to look at environmental or mechanical elements or effects. You can simulate
these without an environmental chamber or shake table.
Then again, just my learned opinion based on experience. :)
SL
Quote from: SolarLab on October 29, 2018, 11:15:29 AM
...
Comsol allows you access to the formulas used in the analysis - you can edit these which is an
extremely huge deal; especially when considering any newly discovered anomalies!
Cost/benefit is pretty good (lots of demo and student deals) and computer requirements are
somewhat viable (or use networked cloud arrays for really complex designs). Overview:
https://www.comsol.com/video/first-look-comsol-multiphysics (https://www.comsol.com/video/first-look-comsol-multiphysics)
...
I knew Comsol by reputation but I am often afraid of how long it takes to master a CAE software. This one seems very complete, so I was doubly afraid and didn't try it. May be I'm wrong.
SolarLab, or someone else, have you used the software and do you have any feedback?
F6FLT, i think you cannot model overunity with a simulator, and i think the reason for that is that for overunity the induced current has to be greater than conventionally given by the rate of change of the magnetic flux.
We cannot know where the overunity is, but we might know where it isn't.
Quote from: ayeaye on November 09, 2018, 05:54:50 AM
F6FLT, i think you cannot model overunity with a simulator...
I know that and said it before. Remember the "
Note to the skeptics" here:
https://overunity.com/17988/ltspice-models-for-bifilar-coils/msg526601/#msg526601
In fact we cannot
search for overunity with a simulator but we can
model it once we know that we get it experimentally and where extra energy comes from (even in LTspice we can input energy in a circuit by putting the appropriate mathematical equation to the appropriate parameter).
Nice to meet you. My name is Kota Imai.
I am a fourth-year university student at Miyazaki University in Japan.
I am not good at English, so I made this text using a translation application. I apologize if it is hard to read.
I am currently researching wireless power transfer systems.
Among them, I am researching on how to improve the transmission distance of wireless power charging systems.
I am thinking of adding a relay circuit to improve the transmission distance, and since there are various previous studies on simply adding a relay circuit, I am aiming to improve the transmission distance by using different types and shapes of coils.
Therefore, I am interested in bifilar coils.
I am planning to add a bifilar coil to the relay circuit only and conduct an experiment to compare it with other coils to see if it improves the transmission distance. However, I do not have much knowledge in electric circuits. So I would like to do a simulation using LTspice before doing the experiment.
I tried to import the bifilar coil elements as described here into LTspice for simulation, but I could not use them in a circuit for a wireless power supply system. So I need you to make an open bifilar coil element in LTspice that can be used in a circuit for a wireless power charging system.
But I am a student and poor so I cannot give you money. But I will introduce you in my graduation thesis which I will present in March 2023. So I would like you to make me an element. Thank you very much.
Please check my posts on my tvkcoil on multisim 8)