I wonder if we should follow along with Luc but use these pancake coils as the transmitter/receiver pair

https://youtu.be/kndYHIHSAE8

edit: video is back online replaced new link

Quote from Matt Watts on July 28th, 2017, 09:00 PM

I wonder if we should follow along with Luc but use these pancake coils as the transmitter/receiver pairs...?

I already tried. But the bifilar windings acts different, due to the dielectric field being between the windings.
With a single wire coil, the top capacitance is the emitter of the dielectric field.
As can be seen in the video, when the dielectric field has a place to flow into (the resonant coil with lamps) the current of the transmitter coil drops down. This is displacement current.
If the displacement current isnt flowing, the current of the transmittter has to push harder, as the dielectric field cant flow properly

if i have a 10nf 400V ac capacitor with 0.1ohm esr

is that the same as 4 series parallel connected equal capacitors? but with a 800V ac rating? still 10nF and 0.1 ohm esr?

High voltage variable (tunable) capacitors.

https://www.youtube.com/watch?v=qR1HN0GZ4IQ

Thank you Mr. Choquette .

Thats a way to do it, but its hard to tune, if the aluminium were movable plates, it could be tuned by changing the surface area.

I added a parallel capacitor to my Phi ratio side by side bifilar speaker wire pancake coil.
the voltage rise goes way up, and also the current draw goes up.

Quote from Steinmetz out of his 1911 book:
In a stationary oscillation of a circuit having uniformly distributed capacity and inductance, That is, the transient of a circuit is storing energy in the dielectric and magnetic field, current and voltage are given by the expression:
then he goes on... read 2 pages in the attached pdf.

A bifilar coil at resonance has these conditions. so the formulas apply to the bifilar coil.

Another quote out of his 1911 book:
Such a transient wave is analogous to the permanent wave of reactive power. As in a stationary (standing) wave, current and voltage ar in quadrature with each other, the question then arises, whether, and what physical meaning a wave has, in which current and voltage are in phase with each other

My god. this is what I have been saying and thinking for a long time. In my words, the resonant voltage rise of the bifilar coil, is due to the magnetic and dielectric fields being in phase, where by the energy of the magnetic field is added up to the dielectric field energy. Because of the big capacity of the the bifilar coil, this explains why the resonant voltage rise can be so big.

In another part of Steinmetz 1911 book, he also talks about b emf, from a coil. explaining the voltage rise. the higher the resistance over the coil the shorter the time, the higher the voltage rise. And stating, just as Eric dollard states, A open circuit(infinite resistance) results in infinit voltage rise, due to the infinite short time.

What my worry is when i want to use Back EMF, is my IGBT is protected from b emf by a diode. So to make it work, a 1:1 transformer should do the trick. The IGBT can pulse one side, while the other side produces a potent back emf, to be shot via diodes into the parallel capacitor of the tuned bifilar coil. the two diodes keep the resonance going in the bifilar coil/capacitor, as they keep the resonance away from the 1:1 transformer.

If it works, the 1:1 bemf coil, should be replaced by a bifilar coil producing b emf, but it cant be switched by a diode protected IGBT, if i want the resirect the B emf into the tuned coil

I think the secret lies in Steinmetz's diagram.

Notice the slope of A, B, C & D.  What I suspect we are stumbling into is the fall-time of our switch (IGBT).  It's like A.  But we want to use it to excite a resonant circuit with a wavelength closer to C.  We need to slow our switch down.  I think this can be done by using a resistor/capacitor RC snubber across the switch.  This will do two things:  Flatten the back EMF spike and make the quarter or half wave resonance match that of the second coil.

maybe Matt, but I dont think so. rather the opposite.
I think using a fast switch without diode makes the following possible:
charge a bifilar coil to a strong magnetic field that creates a strong back emf.

the magnetic strong field is inducing current in the next bifilar coil(carc) and the strong back EMF from the magnetic field collapse is again feeding the capacitor of the carc coil.
the complete energy of the first coil is fed into the carc.

but then the carc is amplifying the current, due to the series capacitance of the bifilar coil and the parallel capacitor being in phase(and being fed by back emf).

then the output coil next to the carc, picks up on the strong amplified magnetic and dielectric field, due to resonance (if tuned properly)

the resonance of the Carc is therefor created by a magnetic field and a dielectric field(from back emf).

so instead of using my igbt with a 1:1 torroid
i want to use a mosfet without protection diode, straight into a bifilar coil so it can produce back emf into the second (carc) coil.

Tried the RC snubber thing and you are correct Evo, that's a dead end.


More experimenting did find an answer you were looking for though.  I pulsed the bifilar coil and dumped all the back EMF into a lamp.  I thought by doing this it would completely zero the output from the collector coil, but it didn't.  I get a small bump.  So we can produce a magnetic field (though small) in addition to collecting the back EMF.

First scopeshot is with no back EMF lamp connected.  This is the normal ringing I get from the collector coil.

Second scopeshot is with the lamp connected.  Notice the bump when the switch opens.

thanks for sharing Matt.
looks like the resonant frequency is around 1.5 MHz am I correct?

I saw the same thing. the back EMF can be tapped, collected and used, without destroying the resonant sine.

but I cant feed the bemf straight into the next coil. maybe I need to convert the spike to a sine first. or like Nelson said,
charge a cap and discharge the cap.

that cap would be the parallel cap from the carc bifilar.

or i need more inductance. I tried to get high votage b emf. but the bifilar quickly saturates. add some ferrite?

or... is the first coil pulsed with b emf from a coil into a cap.... (probably both... add ferrite, use bemf collected in cap to pulse coil, collected bemf again and feed to CARC.

how far could we go with re-using b emf?

We for sure need parametric oscillation, that's a given.  To get it, we need a summator.  What is a summator?

It's an electrical circuit that takes as input two disparate power sources and combines them into a single output that is the sum total of the two inputs.  Think of a big 12 volt lead acid battery combined with a little 1.5 volt penlight battery--you can't just wire them together.

The concept seems simple enough, but is actually more complex than I had initially thought.  You cannot gang together two power sources because the stronger one will always try to reverse power the weaker one.  I learned this when charging two capacitors in series with back EMF.  Doesn't work the way we want at all.

So I'm proposing a possible solution here.  We time slice and multiplex the two power sources.

Suppose we charge two capacitors with the two inputs up to a particular voltage level.  One will charge faster than the other.  When one of them is charged and the other is not, we fire the charged one into a coil/transformer like any common DC-2-DC converter based on some clock signal.  So in this instance, the stronger input power predominately passes through the summator.  So how do we get the weaker input power to contribute?

At some point the weaker power input will finally charge its capacitor up to the preset voltage level we have defined.  When this finally happens, then and only then do we gang the two capacitors together and fire them both, giving us a power burst for that particular cycle.

The beauty of this concept is that the weaker and stronger sources can switch places.  Initially the main power source would be considerably stronger than the feedback power source, but at some point the feedback would actually overtake the main power source--this is where OU operation would begin to take over.

Now that I know we can collect the back EMF while at the same time producing a magnetic field that will induce a current in a collector coil, I'm beginning to see how this could all work together.  If we use the back EMF to drive the CARC coil, then take the power from the collector coil and feed the summator, that's one possible solution and probably the most likely to work.  We will still need some sort of current limiting to put a cap on this feedback loop, otherwise it could take off and self destruct, which it should do left unchecked if we have everything right.

Anyway, I have something here to work on for a while that I'm pretty certain we are going to need in order to make this project a success.  Nelson didn't explicitly tell us about it, let alone how to do it, so I'll work on a solution and put it through some paces to see if my idea will actually work as I envision it.

Quote from Matt Watts on August 5th, 2017, 08:59 AM

We for sure need parametric oscillation, that's a given.  To get it, we need a summator.  What is a summator?

It's an electrical circuit that takes as input two disparate power sources and combines them into a single output that is the sum total of the two inputs.  Think of a big 12 volt lead acid battery combined with a little 1.5 volt penlight battery--you can't just wire them together.

The concept seems simple enough, but is actually more complex than I had initially thought.  You cannot gang together two power sources because the stronger one will always try to reverse power the weaker one.  I learned this when charging two capacitors in series with back EMF.  Doesn't work the way we want at all.

So I'm proposing a possible solution here.  We time slice and multiplex the two power sources.

Suppose we charge two capacitors with the two inputs up to a particular voltage level.  One will charge faster than the other.  When one of them is charged and the other is not, we fire the charged one into a coil/transformer like any common DC-2-DC converter based on some clock signal.  So in this instance, the stronger input power predominately passes through the summator.  So how do we get the weaker input power to contribute?

At some point the weaker power input will finally charge its capacitor up to the preset voltage level we have defined.  When this finally happens, then and only then do we gang the two capacitors together and fire them both, giving us a power burst for that particular cycle.

The beauty of this concept is that the weaker and stronger sources can switch places.  Initially the main power source would be considerably stronger than the feedback power source, but at some point the feedback would actually overtake the main power source--this is where OU operation would begin to take over.

Now that I know we can collect the back EMF while at the same time producing a magnetic field that will induce a current in a collector coil, I'm beginning to see how this could all work together.  If we use the back EMF to drive the CARC coil, then take the power from the collector coil and feed the summator, that's one possible solution and probably the most likely to work.  We will still need some sort of current limiting to put a cap on this feedback loop, otherwise it could take off and self destruct, which it should do left unchecked if we have everything right.

Anyway, I have something here to work on for a while that I'm pretty certain we are going to need in order to make this project a success.  Nelson didn't explicitly tell us about it, let alone how to do it, so I'll work on a solution and put it through some paces to see if my idea will actually work as I envision it.

sounds good matt although I cant completely follow your train of thoughts.

very interesting you say the bemf cant charge a series capacitor bank.

I use series capacitors to get the 10nF.
would it be the bemf only can charge one plate of the cap?

I keep thinking the back emf might be recycled in a resonant circle. back and forth again and again.

with a square wave the current only changes when the pulse is turning high voltage or turning low voltage.

when the voltage is high, there is no change in current. when the voltage is low, there is no  change in current.

only when the voltage changes, the current  changes.

if we use a 1:1 torroid tranformer and use a square wave, the other coil is coupled by the induced magnet field of the toroid.

so the secondary doesnt produce any output when the voltage and current of the primary are stable.

the secondairy shows the large b emf spike when the magnetic field of the toroid collapses.

these spikes have polarity. with a diode we can capture these induced voltage spikes into a capacitor.

if we disconnect this capacitor and reconnect it to a bifilar coil and discharge the capacitor, we dont have the useless dc part of the square wave in the coil. we only have the sharp instant back emf "hammers" ringing the resonant coil next to it

the coil that produces the back emf is not beimg pulsed at its resonant frequency, but at a much lower frequency.
this makes it able to create the b emf spike.

the b emf spike is captured in a capacitor, that is discharged parallel over the next bifilar coil.

but this discharging capacitor forms the resonant circuit with that coil. because of the parallel tuning, the resonant frequency is much lower than the first coil.

so from this coil, there is no back emf spike  because it is tuned into resonance.

this coil not only recieves energy from the parallel capacitor charged with b emf from the first coil, but also from the energy motion of the collapsing field next to it.

until now i used series connected caps. but as Matt Watts pointed out,
you cant capture b emf into (all) the series connected caps.

so i need 1 capacitor to be charged with b emf from the first coil, instead of 22 ...

also. to create bemf i need more inductance on the first coil. add some FERRITE. i dont know where to buy it. so i might grind down some ferrite rods and create some ferrite paint with epoxy resin.

or... smash and grind some magnets, and heat the powder so it looses its magnetism

the field collapse of the first coil creates b emf but the energy motion is also picked ip by the next coil.

when the b emf capacitor is discharged into the next coil forming a tuned resonamt circuit,
the first coil should be charged magneticly again.

this way the both coils get energised.
but can this be done with 1 capacitor?
if we want to keep the coil ringing the parallel cap needs to stay connected.

but to charge the cap with back emf from the first coil, it should be disconnected.

or should it?
maybe a diode is the only thing needed.
to pass the b emf energy into the resonant coil and at the same time keep the resonant coil with the parallel cap separated from the first coil

checked to be sure.
i can charge 2 series capacitors with b emf
so matt. maybe you got a bad connection?

fun thing is the 2 4700uF 16v series caps charged to 48V dc while the scope only showed a 6v back emf spike.

this means the scope simply cant show the fast back emf event properly. its to quick and to high voltage

holy moly
the b emf goes way up when i ground the negative side.
 going to need differnt caps.

lets use those 22 series caps forming 10nF

Quote from evostars on August 6th, 2017, 11:59 AM

i can charge 2 series capacitors with b emf
so matt. maybe you got a bad connection?

Haha.  As long as they are balanced, they'll charge.  Flatten one of them and see if will charge.

I spent hours making sparks, lighting bulbs, playing with this and I'm quite certain what I saw.  The fullest cap wants to take on more charge and the empty cap would rather just be a shunt.

If you are seeing something different, then my back EMF wasn't the same as your back EMF.  That's something we better be sure of it it's true.

Are you collecting the back EMF by way of a diode like I was?

ah that was it. i didnt discharge one of the caps. just a 1:1 ring toroid
one side to igbt switch

other side collecting bemf via diode into cap connected to other side of coil and ground

I'm doing my back EMF tests with the side-by-side coils, power on each coil, switch in between, so that each half of the coil acts like a capacitor plate.  Switch completes the circuit, high voltage lamp across the switch to collect the back EMF when the switch opens.

You can see the very faint glow of the lamp when the switch is not operating.  When the switch is operating, the lamp lights brilliantly.

Quote from Matt Watts on August 6th, 2017, 04:13 PM

I'm doing my back EMF tests with the side-by-side coils, power on each coil, switch in between, so that each half of the coil acts like a capacitor plate.  Switch completes the circuit, high voltage lamp across the switch to collect the back EMF when the switch opens.

You can see the very faint glow of the lamp when the switch is not operating.  When the switch is operating, the lamp lights brilliantly.

thats great! yummy!

brings me to the idea;
collect bemf of 1st coil (open switch1)
into parallel caps of coil 2 (switch 2 open)
close switch 2, discharge caps into coil 2 and let it ring.

switch in the middle like you said Matt.

switch 1 and 2 can synchronous.
i guess another switch for the caps to collect b emf of coil 1 and discharge into coil 2?

Matt what you said reminded me of a quote from nelson rocha:
About your test i'm glad that you could duplicate yourself the " apparently "magnetic  field exclusion with your new configuration ;) nice ! Hope you explore this theme and see if that  are only made by high frequency , or something more happen ;) like stack other pancake coil and look to that like one capacitor exchanging charges between each plate  (two pancake coils ) .

it was in reply to my message stating:   
I have made a car ignition coil, with a radio ferrite rod (so it workes with high frequencies). Making use of the back emf to pulse the bifilar coils.
Works like a charm. no more magnetic field! (finaly fixed it, i wondered what did that in your coils).
 

So... i did so much research i forgot i already had it working with back emf.

now with your knowledge matt, lets see what happens if we feed those coils with back emf, and open them up in the middle.

another quote from me from that message:I also saw tinmans video stating the center connection has more voltage. I tested, and indeed. MUCH more voltage.   

totally forgot it... so glad i found it again.

i probed the center tap with the back emf fed coil and the voltage... spiked.

I looked up that date in my old lab notebook and found this entry:

I  found a radio freq ferrite rod, with a bifilar winding around it. I wound a second bifilar winding around the coil, and connected it to the igbt driver. At low frequencies the current was high, but with high enough frequencies, the current dropped. I used a square wave, at 67% duty. The collapsing field is pushed out via the inside coil. This coil is connected via both wires to the bifilar coil.
The resonant voltage was good, again around 450V sine wave.  Measured with top and bottom coil stacked, outside rim connected, inside rim measured, while the other end is grounded.
I then removed the plastic from the bridge, and I saw the voltage go offscale. I tried to scale down, but is was insanely high. So the bridge does provide the highest voltage in the coil.
Rather spacial! ...
Another special thing. No more magnetic field! BAM