LTspice models for bifilar coils

[gyulasun]

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? 

Anyway, there is an active and friendly yahoo mail group devoted to LTspice, see here:
https://groups.yahoo.com/neo/groups/LTspice/info
Gyula   

[F6FLT]

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.

[gyulasun]

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?

[F6FLT]

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.

[SolarLab]

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

Disclosure: I have nothing what so ever to do with Comsol or anything related - I'm a user only.
Have a great day!

FIN